1
00:00:00,000 --> 00:00:09,760
Welcome to episode 34 of the Language Neuroscience Podcast.

2
00:00:09,760 --> 00:00:13,760
I'm Stephen Wilson and I'm a neuroscientist at the University of Queensland in Brisbane,

3
00:00:13,760 --> 00:00:14,760
Australia.

4
00:00:14,760 --> 00:00:16,960
My guest today is Deborah Levy.

5
00:00:16,960 --> 00:00:20,920
Deb is a language neuroscientist and lecturer in the Princeton Writing Program in Princeton,

6
00:00:20,920 --> 00:00:21,920
New Jersey.

7
00:00:21,920 --> 00:00:25,960
I'm making a few episodes about new papers that catch my eye in the journal Neurobiology

8
00:00:25,960 --> 00:00:26,960
of Language.

9
00:00:26,960 --> 00:00:30,760
This week we're going to talk about Deb's paper, "Role for Left, dorsomedial prefrontal

10
00:00:30,760 --> 00:00:36,320
cortex in self-generated, but not externally cued language production’, which just came out.

11
00:00:36,320 --> 00:00:38,560
It's a lovely paper, as you'll soon see.

12
00:00:38,560 --> 00:00:42,200
I should mention that I actually know Deb very well because she did her PhD in my lab

13
00:00:42,200 --> 00:00:45,720
at Vanderbilt, but we'll try to keep the inside jokes to a minimum.

14
00:00:45,720 --> 00:00:47,400 
Okay, let's get to it.

15
00:00:47,400 --> 00:00:48,640 
Hi, Deb, how are you?

16
00:00:48,640 --> 00:00:49,640 
I'm doing great.

17
00:00:49,640 --> 00:00:50,640 
How are you?

18
00:00:50,640 --> 00:00:51,640 
I'm doing good and 

19
00:00:51,640 --> 00:00:58,240
as we've been talking about, you know, you're in your apartment in West Philly, yeah?

20
00:00:58,240 --> 00:00:59,240 
That's correct, yep.

21
00:00:59,240 --> 00:01:05,840
I'm right by Clark Park and enjoying a nice four o'clock sunshine, I suppose, from

22
00:01:05,840 --> 00:01:06,840 
my apartment.

23
00:01:06,840 --> 00:01:09,680 
Yeah, it looks really, um, idyllic, actually. 

24
00:01:09,680 --> 00:01:13,400
It looks really peaceful and, you know, you've got this loft and like sunshine coming in

25
00:01:13,400 --> 00:01:15,400 
through the windows, so very pleasant.

26
00:01:15,400 --> 00:01:21,120
And I'm just sitting here like in the early dawn hours, like nursing my first coffee that

27
00:01:21,120 --> 00:01:23,120 
really should be a second coffee by this point.

28
00:01:23,120 --> 00:01:25,520 
Yeah, well, it is pretty dreamy here.

29
00:01:25,520 --> 00:01:26,520 
I agree.

30
00:01:26,520 --> 00:01:28,600 
Thank you for complimenting it.

31
00:01:28,600 --> 00:01:31,160 
But your background is very nice as well.

32
00:01:31,160 --> 00:01:32,680 
I like the grey curtain.

33
00:01:32,680 --> 00:01:36,480
The grey curtain is because it's like, this is my office slash laundry.

34
00:01:36,480 --> 00:01:38,720 
And if it wasn't for the grey curtain, you could see my laundry.

35
00:01:38,720 --> 00:01:40,760
In fact, you still actually can see it.

36
00:01:40,760 --> 00:01:42,560 
Very peaceful piece.

37
00:01:42,560 --> 00:01:49,640 
Yeah, but it's like a surprisingly good office for a laundry office.

38
00:01:49,640 --> 00:01:51,920 
We know each other very well.

39
00:01:51,920 --> 00:01:55,960 
You did study in my lab, which was awesome.

40
00:01:55,960 --> 00:01:56,960 
Yes.

41
00:01:56,960 --> 00:02:03,120
But what we're going to talk about today is not that time, but this recent paper that you

42
00:02:03,120 --> 00:02:07,520 
published in neurobiology of language.

43
00:02:07,520 --> 00:02:13,920
Before we get onto that, even though I kind of know this already, can you, for our listeners,

44
00:02:13,920 --> 00:02:18,160
talk about how you got interested in the field of language and the brain?

45
00:02:18,160 --> 00:02:20,640 
Yeah.

46
00:02:20,640 --> 00:02:27,360
When I was very little, I think around the time I learned to read, I was already sort of

47
00:02:27,360 --> 00:02:33,360
really perplexed by the fact that my life had to be filtered through language after that.

48
00:02:33,360 --> 00:02:38,240
So, I remember sitting in the back seat of my parents' car and driving past a billboard

49
00:02:38,240 --> 00:02:42,160
and seeing it and being like, why don't I have the option not to read that?

50
00:02:42,160 --> 00:02:44,560 
I would love I could just look at it.

51
00:02:44,560 --> 00:02:47,000 
But now it's words.

52
00:02:47,000 --> 00:02:52,640
And that sort of mysterious filtering of the world through language continued to fascinate

53
00:02:52,640 --> 00:02:55,520 
me for a very long time.

54
00:02:55,520 --> 00:03:01,080
So, when I was in high school, I got really into Charlie Kaufman movies, and I watched the

55
00:03:01,080 --> 00:03:05,640
Jill Bolte Taylor TED Talk and I was like, I want to spend my whole life thinking about

56
00:03:05,640 --> 00:03:08,280 
how the brain and language interact.

57
00:03:08,280 --> 00:03:12,080 
Just tell us a little bit about that TED Talk for people like them.

58
00:03:12,080 --> 00:03:13,080 
Yeah.

59
00:03:13,080 --> 00:03:18,840
So that is a TED Talk by a neuroscientist who had a stroke in her left hemisphere and she

60
00:03:18,840 --> 00:03:23,720
kind of describes the experience of living her life through her right hemisphere only just

61
00:03:23,720 --> 00:03:28,640 
for the period of time that she was experiencing symptoms.

62
00:03:28,640 --> 00:03:32,960 
And it's really fascinating.

63
00:03:32,960 --> 00:03:37,960
You know, it kind of portrays it as this like meditative, holistic experience of life that

64
00:03:37,960 --> 00:03:43,040
is not kind of dictated by the constraints of language, which, you know, when I was 

65
00:03:43,040 --> 00:03:46,040 
She's very pro-right hemisphere, isn't she?

66
00:03:46,040 --> 00:03:48,800 
I found it like a little bit like scandalous actually.

67
00:03:48,800 --> 00:03:49,800 
Yeah.

68
00:03:49,800 --> 00:03:52,560 
I mean, I was too for a lot of my life.

69
00:03:52,560 --> 00:03:57,800
I think I, you know, longed for a life that wasn't constantly words in my head thinking

70
00:03:57,800 --> 00:03:59,800 
about everything that was going on.

71
00:03:59,800 --> 00:04:00,800 
Wow.

72
00:04:00,800 --> 00:04:02,120 
Yeah.

73
00:04:02,120 --> 00:04:05,920
And then, you know, the more I learned about what the left hemisphere does, the more I was

74
00:04:05,920 --> 00:04:08,200 
like, oh, this guy is pretty important.

75
00:04:08,200 --> 00:04:11,480 
I think, I think I like that I have him around.

76
00:04:11,480 --> 00:04:19,160
Yeah, but anyway, so I think all of those sort of philosophical questions about what does

77
00:04:19,160 --> 00:04:27,000
it mean to live a life through language really kind of like lit my fire about this.

78
00:04:27,000 --> 00:04:32,600
And then, yeah, I went to NYU because they had this major called language in mind, which,

79
00:04:32,600 --> 00:04:35,360 
you know, was very appealing given all of those interests.

80
00:04:35,360 --> 00:04:42,720 
And it was kind of a philosophy psychology linguistics trifecta.

81
00:04:42,720 --> 00:04:46,680
And then when I got to school, I realized that the linguistics and the psych were really

82
00:04:46,680 --> 00:04:48,600 
the things that I was most passionate about.

83
00:04:48,600 --> 00:04:52,760 
So I switched to just a double major in those.

84
00:04:52,760 --> 00:04:57,080 
And I worked in a couple of behavioral labs.

85
00:04:57,080 --> 00:05:03,520
And yeah, it was in a sociolinguistics lab with John Singler, a causal cognition lab with

86
00:05:03,520 --> 00:05:07,000 
Bob Rehder and a visual perception lab with Denis Pelli.

87
00:05:07,000 --> 00:05:11,800
So, I was kind of working in all of those different environments and also doing teaching on

88
00:05:11,800 --> 00:05:12,800 
the side.

89
00:05:12,800 --> 00:05:17,000
So, I did a class called teaching in psychology where I got to TA, the interest site class

90
00:05:17,000 --> 00:05:22,360
for the other undergrads coming in after I took the class, which I loved.

91
00:05:22,360 --> 00:05:25,840 
So yeah, so that was all just quick note.

92
00:05:25,840 --> 00:05:32,160
Like so Denis Pelli is like very secretly famous as one of the coauthors of Psychtoolbox,

93
00:05:32,160 --> 00:05:37,160
which I think, which I think everybody, I mean, many, many of the experience and I feel

94
00:05:37,160 --> 00:05:41,040 
to build on that, on that software, so yeah.

95
00:05:41,040 --> 00:05:43,640 
Yeah, what was that like?

96
00:05:43,640 --> 00:05:44,640 
Well, it was great.

97
00:05:44,640 --> 00:05:46,640 
shoulders of Giants.

98
00:05:46,640 --> 00:05:51,960 
Yeah, I had a lot of fun working in his lab.

99
00:05:51,960 --> 00:05:55,040 
He was, I did my undergraduate thesis in his lab.

100
00:05:55,040 --> 00:06:01,440
So, it was about visual perception of letters at the sort of threshold of visibility and

101
00:06:01,440 --> 00:06:06,040
this experience of those letters kind of like springing into your awareness after just

102
00:06:06,040 --> 00:06:11,240
looking for a long time at something that seems like it isn't there, which we determined

103
00:06:11,240 --> 00:06:12,560 
was sort of a categorical thing.

104
00:06:12,560 --> 00:06:13,720 
Like either it's there, or it isn't.

105
00:06:13,720 --> 00:06:19,960
You kind of don't have any in between experience of them starting to appear.

106
00:06:19,960 --> 00:06:24,000
But he was also working on a project at the time called the Beauty Project that was about

107
00:06:24,000 --> 00:06:26,560 
kind of aesthetic experiences of beauty.

108
00:06:26,560 --> 00:06:32,360
And through the years and I got to, I had a lot of hot takes actually as an undergrad, I

109
00:06:32,360 --> 00:06:34,360 
came in really strong.

110
00:06:34,360 --> 00:06:36,680 
And it didn't totally turn him off of me.

111
00:06:36,680 --> 00:06:40,320 
So yeah, it was a lot of fun.

112
00:06:40,320 --> 00:06:45,520
And I didn't realize that he was so famous for psychtoolbox until I got to your lab and

113
00:06:45,520 --> 00:06:49,880
started using it and saw his name on all of the, you know, all of the documentation.

114
00:06:49,880 --> 00:06:50,880
Yeah.

115
00:06:50,880 --> 00:06:53,880 
And we everyone citing Pellie, 1997.

116
00:06:53,880 --> 00:06:54,880 
Yeah.

117
00:06:54,880 --> 00:06:59,480
And it's kind of really cool how early you got, you knew what your interest was.

118
00:06:59,480 --> 00:07:03,760
I don't know that I've met anybody that actually like underrated enrolled in a major that was

119
00:07:03,760 --> 00:07:05,360 
essentially about language and brain.

120
00:07:05,360 --> 00:07:10,440 
And that's like, you know, surprisingly, you were on your own target.

121
00:07:10,440 --> 00:07:11,440 
Yeah.

122
00:07:11,440 --> 00:07:15,960 
I think that the shape of my interest has morphed quite a bit.

123
00:07:15,960 --> 00:07:16,960 
Yeah.

124
00:07:16,960 --> 00:07:21,440
But the underlying baseline has been really, really consistent since I was about four.

125
00:07:21,440 --> 00:07:23,840 
So that's, that's really cool.

126
00:07:23,840 --> 00:07:24,840 
Yeah.

127
00:07:24,840 --> 00:07:25,840 
Okay.

128
00:07:25,840 --> 00:07:30,600
And so yeah, undergrad, yeah, you did a whole bunch of research as an undergrad.

129
00:07:30,600 --> 00:07:34,600 
And did you go straight into your PhD after that?

130
00:07:34,600 --> 00:07:35,600 
I did not.

131
00:07:35,600 --> 00:07:41,880
I worked in a computational memory lab at Penn for two years as a research assistant.

132
00:07:41,880 --> 00:07:48,120 
So that is the Kahana lab, computational memory lab, studying memory.

133
00:07:48,120 --> 00:07:54,560
And basically, my job there was collecting intracranial data with patients undergoing

134
00:07:54,560 --> 00:07:56,600 
monitoring for epilepsy.

135
00:07:56,600 --> 00:08:01,920
So, I would come in and have them do free recall tasks or, you know, spatial cognition kind

136
00:08:01,920 --> 00:08:07,160
of stuff, you know, set up a little laptop in front of them and encourage them to, you know,

137
00:08:07,160 --> 00:08:10,600
do this while they're hanging out, getting better, really.

138
00:08:10,600 --> 00:08:17,320
And yeah, that was a really different type of experience than I had in undergrad because,

139
00:08:17,320 --> 00:08:20,560
you know, I was going from behavioral psych and behavioral linguistics to this much

140
00:08:20,560 --> 00:08:29,520
more sort of neural computational perspective, which was really cool and really, really mind

141
00:08:29,520 --> 00:08:34,160
boggling for me just to see how many different ways you can look at the same types of questions.

142
00:08:34,160 --> 00:08:37,840 
And I learned so much learned a lot.

143
00:08:37,840 --> 00:08:46,680
And yeah, and then after that is when I decided like all of this neural stuff is really cool,

144
00:08:46,680 --> 00:08:47,880 
but I'm missing the language part.

145
00:08:47,880 --> 00:08:53,200
I really love to do some more language and then, you know, I applied to a very special lab.

146
00:08:53,200 --> 00:08:58,120 
But actually, you applied to Vanderbilt before I was there.

147
00:08:58,120 --> 00:09:01,960 
So I, so why, so like, why did you apply to Vanderbilt?

148
00:09:01,960 --> 00:09:02,960
Because I wasn't there. (Laughter)

149
00:09:02,960 --> 00:09:10,120
Well, I applied because I knew they had a hearing and speech program that was very good.

150
00:09:10,120 --> 00:09:14,920
And I knew that I wanted to do both work that was more kind of focused on language compared

151
00:09:14,920 --> 00:09:17,640 
to what I had been doing as an RA.

152
00:09:17,640 --> 00:09:22,920 
And also work that had some clinical applications.

153
00:09:22,920 --> 00:09:29,680
I think something I've thought about a lot as I've been in my, you know, sort of adult career

154
00:09:29,680 --> 00:09:35,160
is the balance between being interested in something scientifically and being, you know,

155
00:09:35,160 --> 00:09:39,000 
interested in the people that are experiencing what you study.

156
00:09:39,000 --> 00:09:47,560
And I think I was very compelled to feel coupled with the people I was interested in.

157
00:09:47,560 --> 00:09:53,520
And I feel like I was working specifically in areas that would in some way benefit them,

158
00:09:53,520 --> 00:09:55,000 
even if it was long term.

159
00:09:55,000 --> 00:09:56,000
Right.

160
00:09:56,000 --> 00:09:58,120 
So that was a big pull.

161
00:09:58,120 --> 00:09:59,120 
Okay.

162
00:09:59,120 --> 00:10:01,560 
Um, so yeah, that's great.

163
00:10:01,560 --> 00:10:06,040
And then, yeah, so somehow, we got connected and, um, like, yeah, it was through Mike

164
00:10:06,040 --> 00:10:07,040 
de Riesthal.

165
00:10:07,040 --> 00:10:08,040 
I talked to Mike de Riesthal on the phone.

166
00:10:08,040 --> 00:10:11,960
I remember actually sitting in Jefferson hospital while I was on a case talking to Mike

167
00:10:11,960 --> 00:10:16,000 
de Riesthal , like walking around the hallways of Jefferson.

168
00:10:16,000 --> 00:10:17,000 
He seemed great.

169
00:10:17,000 --> 00:10:20,320
He said, you know, based on your interest, there's this guy coming in who you might really

170
00:10:20,320 --> 00:10:25,320
like working with, um, you know, do you want me to just set you guys up on a Skype call

171
00:10:25,320 --> 00:10:26,320 
at the time?

172
00:10:26,320 --> 00:10:27,320 
Probably.

173
00:10:27,320 --> 00:10:28,320 
Yeah.

174
00:10:28,320 --> 00:10:29,320 
Those were the days.

175
00:10:29,320 --> 00:10:30,320 
Yeah.

176
00:10:30,320 --> 00:10:34,560
One thing that I remember from that is that you asked, like, what do I need?

177
00:10:34,560 --> 00:10:37,400 
Like, what should I study before I start?

178
00:10:37,400 --> 00:10:42,160
And then I sent you like a 20 dot point, like syllabus for cognitive neuroscience of

179
00:10:42,160 --> 00:10:43,840 
simple language, which I still have saved.

180
00:10:43,840 --> 00:10:48,280
And it was actually like, that, it's a good document, but it would probably take like about

181
00:10:48,280 --> 00:10:52,680
10 years to get through it all, but like, that makes me feel better about where I am.

182
00:10:52,680 --> 00:10:53,680
But, yeah, I did.

183
00:10:53,680 --> 00:10:55,200 
I mean, I still look back at that too.

184
00:10:55,200 --> 00:10:59,320
I mean, when I try to think like, what are the things that I want to feel like I know

185
00:10:59,320 --> 00:11:02,360 
or that I, you know, know are still ahead of me.

186
00:11:02,360 --> 00:11:04,600 
I use that as sort of a benchmark.

187
00:11:04,600 --> 00:11:05,600 
So great.

188
00:11:05,600 --> 00:11:06,600 
Yeah.

189
00:11:06,600 --> 00:11:09,280 
So two things from your PhD time.

190
00:11:09,280 --> 00:11:15,800
First, can you tell us about what you did with your involvement in the aphasia group at

191
00:11:15,800 --> 00:11:17,040 
at Vanderbilt?

192
00:11:17,040 --> 00:11:18,040 
Yeah.

193
00:11:18,040 --> 00:11:24,240 
So the aphasia group at Vanderbilt is a very, very cool place.

194
00:11:24,240 --> 00:11:30,840
It's run by Dominique Harrington and she every Thursday has people come from really all

195
00:11:30,840 --> 00:11:34,160 
of her Tennessee, but middle Tennessee is kind of the hub.

196
00:11:34,160 --> 00:11:38,000
And although some people commute like three hours to get there, it's very important to

197
00:11:38,000 --> 00:11:39,000 
them.

198
00:11:39,000 --> 00:11:45,960
And it's basically a full day program where there's always kind of conversation and one-to-one

199
00:11:45,960 --> 00:11:49,080 
speech therapy and a real community that's built around that.

200
00:11:49,080 --> 00:11:54,720
So, I came in again, like I mentioned, kind of trying to make sure I stayed connected with

201
00:11:54,720 --> 00:11:58,320 
the people I was interested in, you know, the brains of.

202
00:11:58,320 --> 00:12:04,320 
And I volunteered in that group for basically the duration of my PhD.

203
00:12:04,320 --> 00:12:07,280 
I think it was maybe started the middle of my first year.

204
00:12:07,280 --> 00:12:08,280 
Yeah.

205
00:12:08,280 --> 00:12:14,280
And so, I, you know, because I'm not a clinically trained speech pathologist, I was placed in

206
00:12:14,280 --> 00:12:17,160 
sort of the, we called it the executive group.

207
00:12:17,160 --> 00:12:21,440 
It was very both like relatively mild impairment.

208
00:12:21,440 --> 00:12:26,200
And I was, you know, just kind of hanging out with them and helping, you know, make sure

209
00:12:26,200 --> 00:12:31,720
everybody got birthday cards and, you know, do the, the planning for the group over the

210
00:12:31,720 --> 00:12:33,800 
semester.

211
00:12:33,800 --> 00:12:40,040
But yeah, then Anna Kasdan joined me as well and she was also volunteering in the group and

212
00:12:40,040 --> 00:12:45,560
the two of us decided both that we wanted to sort of capture what was going on there because

213
00:12:45,560 --> 00:12:51,680
it seemed to be so important to the people in the group and also to do our own little spin

214
00:12:51,680 --> 00:12:52,680 
off.

215
00:12:52,680 --> 00:12:56,600
So we had a music and arts group that we started at Vanderbilt, which was so fun.

216
00:12:56,600 --> 00:12:58,600
Anna's a pianist.

217
00:12:58,600 --> 00:13:03,760 
So she ran a choir and I like to think of myself as an amateur artist.

218
00:13:03,760 --> 00:13:07,840
And so, I did, you know, little workshops and I tested them all out to make sure you could

219
00:13:07,840 --> 00:13:12,000 
do them, you know, with one hand at home.

220
00:13:12,000 --> 00:13:15,960
And my husband had a lot of nights of me, you know, like my arm behind my, my back at the

221
00:13:15,960 --> 00:13:18,320 
kitchen table, you know, for me.

222
00:13:18,320 --> 00:13:19,320 
Yeah.

223
00:13:19,320 --> 00:13:22,400 
And so that ended up being really, really fun as well.

224
00:13:22,400 --> 00:13:31,440
So, we have a couple of papers that are out about what that group is and its benefits.

225
00:13:31,440 --> 00:13:39,200
And yeah, I think that was a really important experience for me because it, it got me really

226
00:13:39,200 --> 00:13:44,560
thinking about like what, what the relationship between research and community participation

227
00:13:44,560 --> 00:13:51,520
is, I guess, like research and stakeholders and a lot of the work that I still think about

228
00:13:51,520 --> 00:13:56,000
a lot with a major friendly materials stemmed from that worlds and making sure that there's

229
00:13:56,000 --> 00:13:58,080 
like a way to communicate that information.

230
00:13:58,080 --> 00:14:02,400
Yeah, that's really great that you got to know the people who we were working with in that

231
00:14:02,400 --> 00:14:07,400
way because I just think that it brought so much to the lab that you had that like really

232
00:14:07,400 --> 00:14:14,160
sort of those deep, you know, experiences with actually interacting with people in loads

233
00:14:14,160 --> 00:14:16,280 
of different contexts.

234
00:14:16,280 --> 00:14:20,000
And the other thing I wanted to ask you about is your dissertation, which is then published

235
00:14:20,000 --> 00:14:21,000 
in 2024.

236
00:14:21,000 --> 00:14:25,920 
Can you tell us just briefly about what you worked on for that paper?

237
00:14:25,920 --> 00:14:33,960
Yeah, so that was working with the database that you put together over the course of the

238
00:14:33,960 --> 00:14:38,920
five years, I believe, that you started a Vanderbilt until I wrote the paper.

239
00:14:38,920 --> 00:14:44,040
So that, with all of this speech language pathologists and all of the imaging that was

240
00:14:44,040 --> 00:14:50,680
a mass, we had a really big data set of people with left hemisphere stroke that were tracked

241
00:14:50,680 --> 00:14:53,200 
over the first year of their recovery.

242
00:14:53,200 --> 00:14:58,920
And my dissertation was about trying to predict from the clinical imaging, from what their

243
00:14:58,920 --> 00:15:02,040 
strokes look like on their MRIs.

244
00:15:02,040 --> 00:15:03,040 
Acutely.

245
00:15:03,040 --> 00:15:05,440 
Acutely, yes.

246
00:15:05,440 --> 00:15:09,280
What their language would look like at the one month time point, the three month time point

247
00:15:09,280 --> 00:15:10,920 
and the one year time point.

248
00:15:10,920 --> 00:15:17,760
And that was basically built with support vector regression model, like a machine learning

249
00:15:17,760 --> 00:15:24,840
type approach to predict from the brain images what the language recovery would look like.

250
00:15:24,840 --> 00:15:26,840 
And it did pretty good.

251
00:15:26,840 --> 00:15:28,320 
It did pretty good, didn't it?

252
00:15:28,320 --> 00:15:29,320 
Yeah.

253
00:15:29,320 --> 00:15:35,960
We can predict quite a lot just from the brain, which has a lot of implications, but I guess

254
00:15:35,960 --> 00:15:37,960 
we should get to that another time.

255
00:15:37,960 --> 00:15:38,960 
Yeah.

256
00:15:38,960 --> 00:15:46,840 
So, and then you went to your postdoc with the Chang Lab.

257
00:15:46,840 --> 00:15:49,200
And that is where the paper we're going to talk about today comes from.

258
00:15:49,200 --> 00:15:54,280
So can you tell me about like what it was like to move over to the Chang Lab and get started

259
00:15:54,280 --> 00:15:55,280 
there?

260
00:15:55,280 --> 00:15:56,280 
Yeah.

261
00:15:56,280 --> 00:15:58,640 
Well, it was a big geographical move.

262
00:15:58,640 --> 00:16:03,600
First of all, I moved from Tennessee back to Philadelphia for a month where I got married

263
00:16:03,600 --> 00:16:07,000 
and then immediately moved to San Francisco, like a week after.

264
00:16:07,000 --> 00:16:10,640 
So it was a big physical jump in space.

265
00:16:10,640 --> 00:16:15,120 
But then once I got there, I mean, it was, it was so cool.

266
00:16:15,120 --> 00:16:21,920
It's like an extremely inspiring group of people to work with and really fascinating populations

267
00:16:21,920 --> 00:16:27,320
of people that you can learn about through the neurosurgical resections, which is the

268
00:16:27,320 --> 00:16:29,480 
main data set that I worked with.

269
00:16:29,480 --> 00:16:39,120
So, Eddie is a neurosurgeon and whenever he has left hemisphere cases, he has us do preoperative

270
00:16:39,120 --> 00:16:40,120 
evaluations.

271
00:16:40,120 --> 00:16:45,920
So you see how their language is before surgery, two days after, two to fourish, you see how

272
00:16:45,920 --> 00:16:48,240
they're doing after their surgery.

273
00:16:48,240 --> 00:16:52,640
And then if there's impairment at that point, we follow up a month later and see how their

274
00:16:52,640 --> 00:16:54,160 
language is at that point as well.

275
00:16:54,160 --> 00:16:59,720
So, I got to follow up on very similar types of questions to the dissertation work, but

276
00:16:59,720 --> 00:17:05,840
in a totally different population of people where the, the lesions are, you know, sort of designed

277
00:17:05,840 --> 00:17:11,320 
by a neurosurgeon as opposed to just the, the result of stroke.

278
00:17:11,320 --> 00:17:13,320 
Designed by the MCA.

279
00:17:13,320 --> 00:17:17,640
By the MCA, yeah, although we're not the MCA is actually less relevant for this paper than

280
00:17:17,640 --> 00:17:18,640 
the ACA, but.

281
00:17:18,640 --> 00:17:20,840 
Oh, yeah, that's very true.

282
00:17:20,840 --> 00:17:21,840 
Yeah.

283
00:17:21,840 --> 00:17:22,840
Yeah.

284
00:17:22,840 --> 00:17:26,960
So, in Eddy's lab, would you, what kind of patient interaction this did you have?

285
00:17:26,960 --> 00:17:28,920 
Were you doing the testing and?

286
00:17:28,920 --> 00:17:29,920 
Yeah.

287
00:17:29,920 --> 00:17:33,840
So, I was doing evaluations there, which I was not doing in grad school.

288
00:17:33,840 --> 00:17:35,600 
So that was new for me.

289
00:17:35,600 --> 00:17:41,960
And that I think also added a whole new dimension to, to understanding what assessment is and

290
00:17:41,960 --> 00:17:44,640 
how that, how that bears out.

291
00:17:44,640 --> 00:17:48,800 
Interpersonally, it was fascinating and really.

292
00:17:48,800 --> 00:17:51,800 
Yeah, I learned a lot from, from that as well.

293
00:17:51,800 --> 00:17:57,360
And then for a good chunk of my time at the lab, I was also going to interoperative procedures.

294
00:17:57,360 --> 00:18:04,960
So, I would go into the surgeries and do language tasks during direct cortical stimulation.

295
00:18:04,960 --> 00:18:10,560
So that would be while the, the surgery is happening in order to make sure that it's safe

296
00:18:10,560 --> 00:18:12,320 
to remove certain areas.

297
00:18:12,320 --> 00:18:13,320 
Mm-hmm.

298
00:18:13,320 --> 00:18:17,560
The neurosurgeon stimulates those areas and then tests, you know, can the person still repeat

299
00:18:17,560 --> 00:18:21,800 
or can they still, you know, complete sentences, things like that?

300
00:18:21,800 --> 00:18:26,560
So, Eddie would have been doing the simulations while you were the one administering the language

301
00:18:26,560 --> 00:18:27,560 
stimuli.

302
00:18:27,560 --> 00:18:28,560 
Yeah, yeah, yeah.

303
00:18:28,560 --> 00:18:32,720 
Or documenting it, usually both.

304
00:18:32,720 --> 00:18:38,260
And how long, so I've done, I've been, had the great honor of being in the room one time

305
00:18:38,260 --> 00:18:45,360
throughout one of Eddie's surgeries, which was a seven hour day on a Friday, one day,

306
00:18:45,360 --> 00:18:49,360
or a very memorable experience.

307
00:18:49,360 --> 00:18:50,360 
Is that how it was for you?

308
00:18:50,360 --> 00:18:54,360
Like, you know, it was these very long days, like, um, we were usually not in there for

309
00:18:54,360 --> 00:18:56,120 
the full duration of the surgery.

310
00:18:56,120 --> 00:19:00,520
You would often we'd sort of like huddle in either the sub-sterile room or up in the, the

311
00:19:00,520 --> 00:19:05,240
lab and then kind of try to rush down at the exact moments when it was kind of most useful

312
00:19:05,240 --> 00:19:08,520
for us to be there because I'm not sure if this was your experience, but there's a lot

313
00:19:08,520 --> 00:19:11,520 
more people in the room during surgeries than I would have expected.

314
00:19:11,520 --> 00:19:13,840 
Oh, yeah, there was about 20 people in the room.

315
00:19:13,840 --> 00:19:14,840 
Exactly.

316
00:19:14,840 --> 00:19:19,520
Yeah, so especially when, you know, if you're coming with like a rig or speakers or a microphone

317
00:19:19,520 --> 00:19:22,320
stand, like, you don't want to be in there when you don't have to be because you don't

318
00:19:22,320 --> 00:19:24,560 
want to be in other people's way.

319
00:19:24,560 --> 00:19:29,240
So we, yeah, we would do, I'm, when I say we, I'm talking about me and other people who

320
00:19:29,240 --> 00:19:32,400 
were doing a drop at every search of different sorts.

321
00:19:32,400 --> 00:19:39,560
So, we would often come in around the time that the craniotomy was complete and, you know,

322
00:19:39,560 --> 00:19:49,240
the brain was exposed and then stay for the awake period of the surgery when they sort

323
00:19:49,240 --> 00:19:55,280
of like titrate the anesthetic to have the person be alert for the testing.

324
00:19:55,280 --> 00:19:58,120 
And that'd be about half an hour is right.

325
00:19:58,120 --> 00:20:02,600
Yeah, it would, it would range, but yeah, usually I think around half an hour.

326
00:20:02,600 --> 00:20:09,760
And then sometimes we would stick around for, sometimes the actual like resection procedure

327
00:20:09,760 --> 00:20:13,400 
if there was kind of ongoing monitoring during that.

328
00:20:13,400 --> 00:20:18,040
Or we would, you know, kind of pack up and leave as, you know, efficiently as we can to

329
00:20:18,040 --> 00:20:21,600 
make sure we're, you know, letting the clinician's due their job.

330
00:20:21,600 --> 00:20:26,240
But yeah, I would say we were, we were in and around the operating room for seven hours,

331
00:20:26,240 --> 00:20:30,120
but we were probably only in the operating room for, you know, half an hour to an hour.

332
00:20:30,120 --> 00:20:31,120 
Right.

333
00:20:31,120 --> 00:20:35,200
Yeah, waiting behind the things to come in and do our job and then get out of there.

334
00:20:35,200 --> 00:20:36,200 
Uh-huh.

335
00:20:36,200 --> 00:20:41,040
I was just remembering the one that I, that I was present for about like an hour into

336
00:20:41,040 --> 00:20:42,040 
it.

337
00:20:42,040 --> 00:20:46,680
This, this nurse, I think the senior nurse, like, you know, tapped me on the shoulder

338
00:20:46,680 --> 00:20:51,000
and, and pulled me outside and she was like, can you come outside and like, and then she's

339
00:20:51,000 --> 00:20:54,280 
like, who are you and why are you here? (Laughter)

340
00:20:54,280 --> 00:21:00,040 
And I was like, uh, Eddie invited me.

341
00:21:00,040 --> 00:21:08,360
And she was like, yeah, I also, I had done some operating room stuff at Penn when I was a research

342
00:21:08,360 --> 00:21:09,360 
assistant there.

343
00:21:09,360 --> 00:21:14,400
I will say the people in the operating rooms in San Francisco are very kind and very welcoming

344
00:21:14,400 --> 00:21:16,240 
of the, the vibe in San Francisco.

345
00:21:16,240 --> 00:21:17,240 
That was not my experience.

346
00:21:17,240 --> 00:21:23,080 
But I don't know only for that one, only for that one case, but yeah.

347
00:21:23,080 --> 00:21:28,840
Yeah, maybe I was just, uh, well, I would probably be holding a bunch of electronics so it seemed

348
00:21:28,840 --> 00:21:33,080 
like I had a reason to be there. (Laughter)

349
00:21:33,080 --> 00:21:34,080 
That's funny.

350
00:21:34,080 --> 00:21:35,080
Um, okay.

351
00:21:35,080 --> 00:21:40,080
So yeah, you're very much like embedded in this, in this project and, and doing the data

352
00:21:40,080 --> 00:21:42,080 
collection in many different ways.

353
00:21:42,080 --> 00:21:46,000 
Um, so, uh, let's talk about the paper, right?

354
00:21:46,000 --> 00:21:51,280
So it's called ‘Role for left dorsomedial prefrontal cortex in self-generated, but not externally

355
00:21:51,280 --> 00:21:52,280 
cued language production’.

356
00:21:52,280 --> 00:21:54,280 
Well, you really packed a lot into that title.

357
00:21:54,280 --> 00:21:55,280 
Yeah.

358
00:21:55,280 --> 00:21:56,280 
Yeah.

359
00:21:56,280 --> 00:21:58,480 
And this is in your biology of language 2025.

360
00:21:58,480 --> 00:22:04,840
Just came out and, um, like we talked about, like I'm trying to do some episodes about

361
00:22:04,840 --> 00:22:08,080
papers that are in the journal, because I'm on the editorial board at the journal and I want

362
00:22:08,080 --> 00:22:10,080
it to succeed.

363
00:22:10,080 --> 00:22:15,600
Um, and this paper really struck me as one of the super interesting ones, one of many,

364
00:22:15,600 --> 00:22:19,440 
um, but one that really caught my eye.

365
00:22:19,440 --> 00:22:28,760
Um, so it's about what you call the pre-SMA, um, which is not one of the most popular language

366
00:22:28,760 --> 00:22:29,760 
areas.

367
00:22:29,760 --> 00:22:32,680 
Um, so people might not be so familiar with it.

368
00:22:32,680 --> 00:22:37,440 
So can you start by, um, telling us about what is the pre-SMA?

369
00:22:37,440 --> 00:22:38,440 
Yeah.

370
00:22:38,440 --> 00:22:41,040 
So it's in the medial frontal cortex.

371
00:22:41,040 --> 00:22:46,920
So, I think most images that we see in the neurobiology of language tend to be lateral, you tend

372
00:22:46,920 --> 00:22:50,200
to see like the, the side of the brain with the sylvian fissure everything.

373
00:22:50,200 --> 00:22:56,120
So, if you instead kind of pride the brain open at the longitudinal fissure, I guess you would

374
00:22:56,120 --> 00:23:04,040
see, uh, the medial surface and, um, it's in the, the most kind of, and, well, not the most

375
00:23:04,040 --> 00:23:08,840
anterior part, but the pretty anterior, um, part of the medial prefrontal cortex.

376
00:23:08,840 --> 00:23:14,720
So, um, if you have the SMA that's kind of starting at the central sulcus, um, so that

377
00:23:14,720 --> 00:23:17,000 
stands for supplementary motor area.

378
00:23:17,000 --> 00:23:18,920 
Yes, supplementary motor area.

379
00:23:18,920 --> 00:23:25,480
Um, it's basically in front of that, um, there's a delineation called the, um, the VCA line

380
00:23:25,480 --> 00:23:31,960
that kind of distinguishes them that's aligned with the anterior commissure and that area in

381
00:23:31,960 --> 00:23:35,160 
front of the SMA is what's referred to as the pre-SMA.

382
00:23:35,160 --> 00:23:37,360 
So the pre-Supplementary motor area.

383
00:23:37,360 --> 00:23:38,360 
Okay.

384
00:23:38,360 --> 00:23:42,200
Um, and can you kind of, so you've kind of explained the, the relative, uh, some of the

385
00:23:42,200 --> 00:23:48,080
call location with respect to the SMA proper, um, what, and what, prior to your work, like,

386
00:23:48,080 --> 00:23:53,440
coming into it, like, what was generally known about the functional role of the SMA versus

387
00:23:53,440 --> 00:23:56,040 
the pre-SMA, just kind of big picture?

388
00:23:56,040 --> 00:23:57,040 
Yeah.

389
00:23:57,040 --> 00:24:04,200
So, generally, I think the SMA is more associated with, like, motor output, um, so some kind

390
00:24:04,200 --> 00:24:11,160
of, you know, connections with motor cortex and motor planning, um, things where if you

391
00:24:11,160 --> 00:24:16,520
were to lesion it, you would get, um, deficits in kind of movement or, or motor stuff.

392
00:24:16,520 --> 00:24:21,960
Um, whereas the pre-SMA seems to be more kind of abstract in what it works with.

393
00:24:21,960 --> 00:24:26,360 
So it's more associated with, um, prefrontal cortex.

394
00:24:26,360 --> 00:24:29,000 
It's less motor in terms of its connections.

395
00:24:29,000 --> 00:24:36,440
It, uh, seems to respond to less specific types of, um, tasks, uh, or, or different types

396
00:24:36,440 --> 00:24:37,920 
of tasks, I should say.

397
00:24:37,920 --> 00:24:42,600
Uh, it seems to be working at a level that's not quite as directly tied to motor output

398
00:24:42,600 --> 00:24:47,040 
and is more about either planning or decision making related to those.

399
00:24:47,040 --> 00:24:48,040 
Okay.

400
00:24:48,040 --> 00:24:49,040 
Great.

401
00:24:49,040 --> 00:24:54,760
Um, and, you know, I think the first person, I'm not sure if he was first, but like, the

402
00:24:54,760 --> 00:25:00,840
first prominent person to sort of relate these areas to language was Wilder Penfield.

403
00:25:00,840 --> 00:25:05,680
Um, do you think that he would, I mean, and was, do you think he was talking about SMA or

404
00:25:05,680 --> 00:25:08,080 
pre-SMA in his, um, work?

405
00:25:08,080 --> 00:25:09,840 
That's a good question.

406
00:25:09,840 --> 00:25:13,880
I'm trying, I, I remember an image that you're referring to, like, of the medial surface

407
00:25:13,880 --> 00:25:15,840 
with the hatching over that area.

408
00:25:15,840 --> 00:25:16,840 
Yeah.

409
00:25:16,840 --> 00:25:17,840 
Um, yeah.

410
00:25:17,840 --> 00:25:23,800
I, if I remember correctly, the, the, the pre-SMA wasn't like given its own name

411
00:25:23,800 --> 00:25:25,960 
until a little bit later than that.

412
00:25:25,960 --> 00:25:27,840 
Am I, do you agree?

413
00:25:27,840 --> 00:25:28,840 
I think I do.

414
00:25:28,840 --> 00:25:33,880
I mean, I, I'm not, um, I think he was talking about SMA, like, yeah, I think so.

415
00:25:33,880 --> 00:25:38,120
Um, and, and, you know, he noticed that when you, like, it was one of the three brain regions

416
00:25:38,120 --> 00:25:44,160
where if you stimulated it, you could cause, um, you know, kind of transient, um, aphasia

417
00:25:44,160 --> 00:25:45,160 
or speech arrest, really.

418
00:25:45,160 --> 00:25:49,560
I mean, it's not strictly, not necessarily aphasia, but, yeah, speech arrest.

419
00:25:49,560 --> 00:25:50,560 
So, yeah.

420
00:25:50,560 --> 00:25:55,040
You know, but I think, yeah, he's probably like really hitting on that motor, more posterior

421
00:25:55,040 --> 00:25:59,680 
area, whereas you guys are talking about the area in front of that.

422
00:25:59,680 --> 00:26:04,960
Um, and so that area that's in front of, like, so the pre-SMA, like, is there much prior

423
00:26:04,960 --> 00:26:10,120
to your papers, there much in the language literature about that region?

424
00:26:10,120 --> 00:26:15,800
I have a couple of thoughts on that, uh, trans-cortical motor aphasia is, you know, like 1885,

425
00:26:15,800 --> 00:26:18,040 
1886 is when it was first described.

426
00:26:18,040 --> 00:26:20,520 
Um, and, but not with an anatomical.

427
00:26:20,520 --> 00:26:21,520 
It's an anatomical basis.

428
00:26:21,520 --> 00:26:22,520 
No, it's an anatomical basis.

429
00:26:22,520 --> 00:26:23,520
And in the time.

430
00:26:23,520 --> 00:26:24,520 
Yeah.

431
00:26:24,520 --> 00:26:30,040
Um, I think the, the neuropsychology literature talks about this area a little bit more, um,

432
00:26:30,040 --> 00:26:35,760
when it comes to superior, um, medial front lesions that lead to what people might have

433
00:26:35,760 --> 00:26:41,480
called like a dis-executive syndrome or some kind of, uh, issue with, um, executive functioning

434
00:26:41,480 --> 00:26:46,840
that kind of ties into language, but isn't necessarily, um, specifically part of, like,

435
00:26:46,840 --> 00:26:47,840 
the language apparatus.

436
00:26:47,840 --> 00:26:50,840 
Mm-hmm.

437
00:26:50,840 --> 00:26:54,160 
So it grew up there, um, it also, I think, do you still--

438
00:26:54,160 --> 00:26:57,920
And you cited a lot of papers by my Queensland colleague, Gail Robinson?

439
00:26:57,920 --> 00:26:58,920 
Yes, I did.

440
00:26:58,920 --> 00:27:05,600
But, yeah, and she writes a lot about kind of distinctions between the, uh, medial frontal

441
00:27:05,600 --> 00:27:10,160
areas and lateral frontal areas, uh, and they're possible differential contributions to

442
00:27:10,160 --> 00:27:13,480 
different types of language, uh, dysfunction.

443
00:27:13,480 --> 00:27:18,840
And, yeah, so she's written lots and lots of case studies about, um, people either after

444
00:27:18,840 --> 00:27:25,240
stroke or after a tumor resection, um, sometimes there's neurodegenerative conditions like PSP or

445
00:27:25,240 --> 00:27:27,080 
or Parkinson's where this shows up.

446
00:27:27,080 --> 00:27:33,200
Um, and it's very tied to what she would call an energization, uh, deficit or what 

447
00:27:33,200 --> 00:27:38,640
Alexander would have called an energization deficit as well, um, which is kind of about

448
00:27:38,640 --> 00:27:42,640 
both initiating and sustaining a response over time.

449
00:27:42,640 --> 00:27:49,840
Um, that seems to be potentially domain general, not necessarily language specific, but certainly

450
00:27:49,840 --> 00:27:55,200
bears out in language and seems to be very tightly associated with those medial frontal

451
00:27:55,200 --> 00:27:56,200 
areas.

452
00:27:56,200 --> 00:28:02,520
Um, and then there's, Luria wrote about dynamic aphasia, which is what, you know, we kind

453
00:28:02,520 --> 00:28:10,920
of end up mapping this concept onto the, in terms of the profile, uh, in the 1940s, I believe,

454
00:28:10,920 --> 00:28:16,480 
and, Luria, my impression, I'd love to hear your thoughts on this.

455
00:28:16,480 --> 00:28:20,720
My impression is again that he's much more discussed among neuropsychologists than among,

456
00:28:20,720 --> 00:28:25,400 
like, modern day study years of the neurobiology of language.

457
00:28:25,400 --> 00:28:26,920 
Um, what are your thoughts?

458
00:28:26,920 --> 00:28:31,360 
Oh, yeah, I think he's, um, like, criminally neglected by our field.

459
00:28:31,360 --> 00:28:32,360 
Yeah.

460
00:28:32,360 --> 00:28:33,360 
Yeah.

461
00:28:33,360 --> 00:28:34,360 
Yeah.

462
00:28:34,360 --> 00:28:37,480
It's just like it's, it's very like, you know, just this whole stream of aphasiology

463
00:28:37,480 --> 00:28:41,240 
that we kind of ignore that had like a lot of insights.

464
00:28:41,240 --> 00:28:42,240 
Yeah.

465
00:28:42,240 --> 00:28:46,880
And it's always really interesting to kind of try and understand like how all of his concepts

466
00:28:46,880 --> 00:28:49,280 
mapped on to like Western concepts.

467
00:28:49,280 --> 00:28:50,280 
Right.

468
00:28:50,280 --> 00:28:51,280 
Yeah.

469
00:28:51,280 --> 00:28:54,400
And I think there's a reasonably good correspondence between dynamic aphasia and transcortical

470
00:28:54,400 --> 00:28:58,520
motor aphasia, but like there's subtle differences in the understanding.

471
00:28:58,520 --> 00:28:59,520 
Yeah.

472
00:28:59,520 --> 00:29:00,520 
Right.

473
00:29:00,520 --> 00:29:01,520 
Yeah.

474
00:29:01,520 --> 00:29:05,400
Um, so that is good for me to hear coming from you because I, when I started reading

475
00:29:05,400 --> 00:29:07,800 
Luria, I was like, oh my gosh, this is also relevant.

476
00:29:07,800 --> 00:29:09,240 
Why didn't I know this?

477
00:29:09,240 --> 00:29:13,000 
And I think it might just be because have to blame your teachers.

478
00:29:13,000 --> 00:29:15,000 
No, I wouldn't think like teachers.

479
00:29:15,000 --> 00:29:17,800
I would blame the fact that like Russia wasn't part of the United States.

480
00:29:17,800 --> 00:29:20,600 
And so there was like just a different research tradition.

481
00:29:20,600 --> 00:29:24,320
And, um, I think some of his stuff didn't even get translated until the 70s.

482
00:29:24,320 --> 00:29:30,160
And there was already, you know, some dominant, uh, you know, schools of thinking going

483
00:29:30,160 --> 00:29:32,880 
on with, you know, the Boston school and everything.

484
00:29:32,880 --> 00:29:39,840
So it kind of didn't match back up, but anyway, so he described dynamic aphasia back in the

485
00:29:39,840 --> 00:29:47,520
40s as, you know, pretty clearly this, uh, difference between language as kind of a function

486
00:29:47,520 --> 00:29:52,000
and language and practice when it comes to spontaneous speech or like this propositional

487
00:29:52,000 --> 00:29:53,240 
part of language.

488
00:29:53,240 --> 00:29:58,000
And, uh, that's pretty much exactly what we were observing in these patients that we're

489
00:29:58,000 --> 00:29:59,000 
going to talk about soon.

490
00:29:59,000 --> 00:30:00,000 
Yeah.

491
00:30:00,000 --> 00:30:01,000 
Okay.

492
00:30:01,000 --> 00:30:03,000 
I wasn't familiar with the term.

493
00:30:03,000 --> 00:30:08,600
Um, so it was a lot of kind of learning backwards after I had observed something, uh, that

494
00:30:08,600 --> 00:30:11,080 
it had actually existed for quite some time.

495
00:30:11,080 --> 00:30:12,080 
Yeah.

496
00:30:12,080 --> 00:30:17,640
Um, and then another person that, um, you cite that's talks about this area is Jeff

497
00:30:17,640 --> 00:30:24,160
Binder, um, with, um, in particular, his 2009, um, meta-analysis, which I find to be a

498
00:30:24,160 --> 00:30:25,600 
really useful paper.

499
00:30:25,600 --> 00:30:27,800 
Um, so what does he say about this area?

500
00:30:27,800 --> 00:30:31,880 
Yeah, I believe he says it's involved in sort of semantic selection.

501
00:30:31,880 --> 00:30:34,040 
Um, this is from the meta-analysis, right?

502
00:30:34,040 --> 00:30:37,760
So that's, you know, all these different studies about semantics and, you know, various

503
00:30:37,760 --> 00:30:42,960
different task forms, uh, and that mediocre, prefrontal area does show up.

504
00:30:42,960 --> 00:30:49,360
Um, and he also mentions that it's kind of quite overlooked in language literature overall.

505
00:30:49,360 --> 00:30:55,160
Um, but, yeah, I think it's about retrieving semantic concepts and kind of, um, deciding

506
00:30:55,160 --> 00:31:00,920
which ones are potentially relevant for the situation at hand to, to use.

507
00:31:00,920 --> 00:31:02,240
Okay, great.

508
00:31:02,240 --> 00:31:07,520
Um, yes, so there's like kind of definitely, um, a back drop in the literature, like even

509
00:31:07,520 --> 00:31:11,800
though this is not an area that's kind of, you know, usually talked about as being one

510
00:31:11,800 --> 00:31:15,720
of the major language areas, there's definitely like data out there that suggests that it

511
00:31:15,720 --> 00:31:17,960 
does play some kind of a role.

512
00:31:17,960 --> 00:31:24,760
Um, and then you, um, find yourself with, um, this very unique patient population.

513
00:31:24,760 --> 00:31:29,120
Um, and you notice some things with patients who have lesions to this area.

514
00:31:29,120 --> 00:31:33,240
So in, in your paper, you start the way that you presented it, I think is very effective,

515
00:31:33,240 --> 00:31:37,360
which is that you start with this case, vignette before you get into all the analyses.

516
00:31:37,360 --> 00:31:40,360
Um, so I thought that might also be a good way for us to talk about it.

517
00:31:40,360 --> 00:31:48,760
So can you tell me about the, the case of, um, EK, I guess, um, who's not anonymous, um,

518
00:31:48,760 --> 00:31:52,520 
and just tell me about what, what you, um, observed with, with him.

519
00:31:52,520 --> 00:31:53,520 
Yeah.

520
00:31:53,520 --> 00:31:58,000
Um, as a side note, I actually had initially written it as something like EK or Dr K,

521
00:31:58,000 --> 00:32:00,880
and he specifically wrote to me and asked like, can you just use my name?

522
00:32:00,880 --> 00:32:02,400 
This makes me feel really weird.

523
00:32:02,400 --> 00:32:06,320 
So, so if it's okay with you, I'll go with Edwin.

524
00:32:06,320 --> 00:32:08,960 
Yeah, yeah, you can go with it, however you like.

525
00:32:08,960 --> 00:32:09,960 
Yeah.

526
00:32:09,960 --> 00:32:16,480
Um, but yeah, so, uh, I met Edwin before his surgery in the pre-op appointment, um, and

527
00:32:16,480 --> 00:32:22,720
he was multi-lingual, no, he spoke Cantonese as well as English, possibly some other languages

528
00:32:22,720 --> 00:32:23,720 
as well.

529
00:32:23,720 --> 00:32:29,400
Um, so we were talking during the evaluation, we used the quick aphasia battery, and afterwards

530
00:32:29,400 --> 00:32:34,480
we did a little brief multi-lingualism screener and he had a lot of opinions about it, um,

531
00:32:34,480 --> 00:32:40,800
because his work, he's a, uh, at the time was a student studying linguistics at the PhD

532
00:32:40,800 --> 00:32:44,080 
level, um, specifically like language revitalization.

533
00:32:44,080 --> 00:32:48,240
So, he had a lot of, you know, opinions, and I remember just thinking it was really cool

534
00:32:48,240 --> 00:32:50,200 
and interesting to talk to him.

535
00:32:50,200 --> 00:32:58,200
And, you know, we were talking at a, a pretty high level about language and, um, yeah, and

536
00:32:58,200 --> 00:33:04,360
then two days later, I go to see him post-op and I was sometimes, I think, so his language

537
00:33:04,360 --> 00:33:06,000 
would have been normal pre-op?

538
00:33:06,000 --> 00:33:10,280
Yes, it, yeah, I, I believe just 10 out of 10 flying colors across the board.

539
00:33:10,280 --> 00:33:14,280
And, and, and so, well, yes, so why is he getting a chunk of his brain taken out?

540
00:33:14,280 --> 00:33:23,440
Yeah, so he had, uh, Astrocytoma in his medial frontal cortex, uh, which they, his life

541
00:33:23,440 --> 00:33:29,800
was going on pretty standardly until he was on a Zoom call with his PhD advisor actually,

542
00:33:29,800 --> 00:33:33,960
um, and had a seizure in the middle of the Zoom call, uh, which I believe was his first

543
00:33:33,960 --> 00:33:34,960 
ever seizure.

544
00:33:34,960 --> 00:33:38,200 
Um, first and I think, I think from your description.

545
00:33:38,200 --> 00:33:39,200 
I think that's true.

546
00:33:39,200 --> 00:33:40,200 
Yeah.

547
00:33:40,200 --> 00:33:48,840
Um, and, uh, yes, so then he went to, he was already in the San Francisco area, um, went to

548
00:33:48,840 --> 00:33:51,760 
UCSF and, you know, they found a mass on imaging.

549
00:33:51,760 --> 00:33:57,840
So he's going in for, um, resective surgery to have that area removed, essentially to have

550
00:33:57,840 --> 00:33:58,840 
the tumor taken out.

551
00:33:58,840 --> 00:33:59,840
Okay.

552
00:33:59,840 --> 00:34:05,200
So, yeah, so pre-op is where I met him the day before that appointment.

553
00:34:05,200 --> 00:34:10,960
Um, and we do language assessments just to make sure everything is, uh, we have a basis of

554
00:34:10,960 --> 00:34:14,400 
comparison really for what we see after the surgery.

555
00:34:14,400 --> 00:34:19,200
And yeah, I believe just tens across the board on the quick aphasia battery.

556
00:34:19,200 --> 00:34:24,760
I think he had described some word finding issues, but, uh, we were not picking them up with

557
00:34:24,760 --> 00:34:25,760 
the quick aphasia battery.

558
00:34:25,760 --> 00:34:28,920 
I think he was largely within normal limits.

559
00:34:28,920 --> 00:34:29,920 
Mm-hmm.

560
00:34:29,920 --> 00:34:31,160 
So that was pretty up.

561
00:34:31,160 --> 00:34:32,920 
Um, hmm.

562
00:34:32,920 --> 00:34:39,960
And then post-op I, I guess we should say is the, what was taken at, what was resected?

563
00:34:39,960 --> 00:34:43,960
And there's a very nice picture, there's a very nice picture of it in the, you know, one

564
00:34:43,960 --> 00:34:47,360 
that this, this is what you, this is an audio format.

565
00:34:47,360 --> 00:34:49,360 
So can you describe?

566
00:34:49,360 --> 00:34:54,480
Um, well, essentially the area I was describing before, just in front of the SMA, uh, the

567
00:34:54,480 --> 00:34:58,880 
pre-SMA is what is removed in, um, in this patient Edwin.

568
00:34:58,880 --> 00:35:03,560
So, um, it's pretty, I would say pretty specific to that area, it doesn't go too much into anterior

569
00:35:03,560 --> 00:35:06,080 
interior, it doesn't go too much inferior.

570
00:35:06,080 --> 00:35:11,160
It's pretty much just, I would say our region of interest, um, although it became interest

571
00:35:11,160 --> 00:35:12,640 
after the fact.

572
00:35:12,640 --> 00:35:13,640 
Yes, exactly.

573
00:35:13,640 --> 00:35:14,640 
Um, okay.

574
00:35:14,640 --> 00:35:20,480
And then how does he do, um, sorry, my dog's decided that she needs to enter this room,

575
00:35:20,480 --> 00:35:22,280 
but that's not going to happen.

576
00:35:22,280 --> 00:35:27,520
Um, what, so then what does he, uh, how does he look when you meet him again a few days

577
00:35:27,520 --> 00:35:28,520 
later?

578
00:35:28,520 --> 00:35:33,240
So when I first walked into the room, I was expecting maybe we'd have a pretty similar

579
00:35:33,240 --> 00:35:35,800 
interaction to the one we had beforehand.

580
00:35:35,800 --> 00:35:39,800
Um, sometimes I would sort of make predictions like, okay, based on the reception location, what

581
00:35:39,800 --> 00:35:40,800 
would I expect?

582
00:35:40,800 --> 00:35:45,840
Uh, and I didn't really have a ton of expectations about this particular area.

583
00:35:45,840 --> 00:35:51,320
So, um, and oftentimes we would see people after surgery and, you know, in about 30% of cases,

584
00:35:51,320 --> 00:35:52,760 
they're still within normal limits.

585
00:35:52,760 --> 00:35:55,560 
So I was thinking maybe that would be the case with him.

586
00:35:55,560 --> 00:35:58,080 
Uh, um, I'm only first started interacting.

587
00:35:58,080 --> 00:36:00,320 
I was still under that impression.

588
00:36:00,320 --> 00:36:01,640 
I said, like, hey, how are you?

589
00:36:01,640 --> 00:36:03,320 
You said, good, good to see you.

590
00:36:03,320 --> 00:36:05,440 
You know, that kind of, that kind of thing.

591
00:36:05,440 --> 00:36:10,200
Um, but then we start going into the quick aphasia battery, which one of the first tasks

592
00:36:10,200 --> 00:36:14,560
is this connected speech task, which is essentially just, you know, asking someone to speak

593
00:36:14,560 --> 00:36:18,040 
spontaneously about some experience in their life.

594
00:36:18,040 --> 00:36:21,880
In this case, it might be the surgery or it might be like a story about something from

595
00:36:21,880 --> 00:36:23,080 
their past.

596
00:36:23,080 --> 00:36:29,920
Um, and I noticed pretty quickly that he was almost not communicating at all.

597
00:36:29,920 --> 00:36:37,040
Um, he was trying to and he had, uh, you know, he would respond kind of with the first bit

598
00:36:37,040 --> 00:36:38,560 
of an answer to a question.

599
00:36:38,560 --> 00:36:40,480 
So I would say, like, what do you remember from surgery?

600
00:36:40,480 --> 00:36:47,400
And he'd say, well, I remember waking up and then it would be long pauses.

601
00:36:47,400 --> 00:36:52,880
And then I think I did a picture association task, you know, this very kind of like,

602
00:36:52,880 --> 00:36:56,920 
technical language is still, um, easily being retrieved.

603
00:36:56,920 --> 00:37:05,160
Um, but at a rate and, uh, you know, a sort of degree of unusual pausing that you would

604
00:37:05,160 --> 00:37:07,880 
not expect from somebody without a language impairment.

605
00:37:07,880 --> 00:37:12,440
So at that point, I thought, okay, so probably the rest of the evaluation is going to go much

606
00:37:12,440 --> 00:37:13,440 
more like this.

607
00:37:13,440 --> 00:37:18,000
Like I think he's probably quite anomic, um, I'm expecting there will be difficulties

608
00:37:18,000 --> 00:37:20,720 
naming pictures, difficulties describing things.

609
00:37:20,720 --> 00:37:22,720 
Um, yeah.

610
00:37:22,720 --> 00:37:28,280
So his conversational speech was like super sparse and just like he couldn't really generate anything

611
00:37:28,280 --> 00:37:29,940 
in answer to your questions.

612
00:37:29,940 --> 00:37:30,940 
Right.

613
00:37:30,940 --> 00:37:31,940 
Exactly.

614
00:37:31,940 --> 00:37:32,940 
All of them.

615
00:37:32,940 --> 00:37:40,600 
Um, um, um, um, um, um, um, um, uh, yeah, exactly.

616
00:37:40,600 --> 00:37:44,280 
Um, and it, it didn't seem like it was for any lack of trying.

617
00:37:44,280 --> 00:37:48,940
I mean, he, he was definitely working towards communicating with me, you know, making efforts

618
00:37:48,940 --> 00:37:54,900
to communicate with me, but just kind of nothing, nothing coming up to, you know, to be described

619
00:37:54,900 --> 00:37:55,900
in words.

620
00:37:55,900 --> 00:38:00,860
Um, but on the rest of the evaluation, he actually still did pretty much perfectly.

621
00:38:00,860 --> 00:38:03,220 
So he could name every picture.

622
00:38:03,220 --> 00:38:12,620
Uh, he could, um, describe the, the images of, you know, that illicit grammatical constructions.

623
00:38:12,620 --> 00:38:15,180 
He could do the motor speech tasks.

624
00:38:15,180 --> 00:38:19,300
He could say catastrophe, catastrophe, catastrophe, which is like a notoriously difficult word

625
00:38:19,300 --> 00:38:21,140 
if you have motor speech issues.

626
00:38:21,140 --> 00:38:27,340
Um, and, you know, if I had only seen the last two thirds of the evaluation, I'd probably

627
00:38:27,340 --> 00:38:32,780
say this person has normal language, uh, but it was really just that spontaneously generated

628
00:38:32,780 --> 00:38:36,260 
portion that was completely different.

629
00:38:36,260 --> 00:38:39,980 
And that was very striking to me.

630
00:38:39,980 --> 00:38:40,980
So yeah.

631
00:38:40,980 --> 00:38:46,940
And by the way, I just want to say like, um, listeners should check out the paper because

632
00:38:46,940 --> 00:38:52,420
your figure one A is like an extremely nice figure because it kind of encapsulates everything

633
00:38:52,420 --> 00:38:55,780 
about this paper in one neat little panel.

634
00:38:55,780 --> 00:39:00,340
It's got a picture of Edwin's brain with the missing region, like with the nice dotted

635
00:39:00,340 --> 00:39:01,340 
outline around it.

636
00:39:01,340 --> 00:39:03,980 
And it's got four sections from the quick aphasia battery.

637
00:39:03,980 --> 00:39:08,460
One is repetition and you get to see him repeating like a really complicated sentence.

638
00:39:08,460 --> 00:39:12,620
And these confrontation naming you get to see like these rather tricky low frequency

639
00:39:12,620 --> 00:39:17,780
lexical items that he retrieves speech motor, you get to see him saying catastrophe, catastrophe.

640
00:39:17,780 --> 00:39:21,420
And then the other corner is self-generated speech and it's like, what do you remember from

641
00:39:21,420 --> 00:39:22,500
the surgery?

642
00:39:22,500 --> 00:39:32,460
Um, so I work up, um, I, um, size, um, and it just really makes that like dissociation super

643
00:39:32,460 --> 00:39:35,420 
clear, um, and encapsulates.

644
00:39:35,420 --> 00:39:36,900 
So nice, nice figure.

645
00:39:36,900 --> 00:39:37,900 
It really should check it out.

646
00:39:37,900 --> 00:39:41,380 
Um, now what were you going to say next?

647
00:39:41,380 --> 00:39:51,500
Um, yeah, I guess I would say that well, one of the first things I thought is this is surprising,

648
00:39:51,500 --> 00:39:55,540
but I actually, it seems a little familiar to me and I remembered somebody I'd seen earlier

649
00:39:55,540 --> 00:40:02,500
that year who actually had sort of a similar presentation, um, after surgery, although slightly

650
00:40:02,500 --> 00:40:09,420 
less, um, kind of involved in the, the assessment.

651
00:40:09,420 --> 00:40:15,940
So, I remembered her evaluation and then I looked at her imaging and I was like, oh, that was

652
00:40:15,940 --> 00:40:16,940 
the same place.

653
00:40:16,940 --> 00:40:19,980 
Actually, I think that was almost exactly the same location.

654
00:40:19,980 --> 00:40:26,060
Um, so it definitely kind of like sparked a, an awareness in me that this might be a pattern

655
00:40:26,060 --> 00:40:28,780 
and not just an interesting case.

656
00:40:28,780 --> 00:40:35,500
Uh, uh, and then I did a follow up evaluation with, uh, Edwin a month later and he was essentially

657
00:40:35,500 --> 00:40:36,900 
back to his baseline.

658
00:40:36,900 --> 00:40:43,700
Um, so he could speak in great detail about the experience he had doing that evaluation

659
00:40:43,700 --> 00:40:48,300
and kind of realizing that he had difficulty communicating even though it didn't really feel

660
00:40:48,300 --> 00:40:50,140 
internally like he should.

661
00:40:50,140 --> 00:40:55,620
Um, so he gave a lot of really interesting insights, which he then brought up as a supplementary,

662
00:40:55,620 --> 00:40:57,740 
um, material for this paper.

663
00:40:57,740 --> 00:41:04,740
Um, that I thought were for a variety of reasons just shed so much light on what it's actually

664
00:41:04,740 --> 00:41:08,300 
like from the perspective of the person experiencing this.

665
00:41:08,300 --> 00:41:12,500
And, uh, yeah, I felt very strongly that he should be involved in, in writing this paper

666
00:41:12,500 --> 00:41:16,580 
up because it was so, you have to have active.

667
00:41:16,580 --> 00:41:19,540 
And so yeah, you've included him as a co-author and I love that.

668
00:41:19,540 --> 00:41:23,940
Like, you know, that, I mean, that's really like bringing the people that, whose brains we

669
00:41:23,940 --> 00:41:30,300
study into the process, um, even more richly than, you know, we would normally be like putting

670
00:41:30,300 --> 00:41:32,180 
them on advisory boards and stuff.

671
00:41:32,180 --> 00:41:36,660
I mean, I love the idea of like, you know, one of the participants being a co-author.

672
00:41:36,660 --> 00:41:42,060
I mean, and what would you say was the most like, what were the biggest insights that you

673
00:41:42,060 --> 00:41:47,580
got from his description of what it had been like to be in this, um, ultimately transient

674
00:41:47,580 --> 00:41:49,580 
aphasia?

675
00:41:49,580 --> 00:41:50,580 
Yeah.

676
00:41:50,580 --> 00:41:53,380 
I think two things stuck out to me.

677
00:41:53,380 --> 00:41:59,580
One was that metaphor about, uh, feeling, I should I direct readers to this metaphor in

678
00:41:59,580 --> 00:42:01,180 
the paper or should I read it out loud?

679
00:42:01,180 --> 00:42:02,980 
It says, uh, yeah, essentially.

680
00:42:02,980 --> 00:42:03,980 
Would it be a life?

681
00:42:03,980 --> 00:42:04,980 
Yeah.

682
00:42:04,980 --> 00:42:12,180
So, he has this metaphor about, um, trying to describe what it was like to retrieve the

683
00:42:12,180 --> 00:42:13,660 
words that he couldn't get.

684
00:42:13,660 --> 00:42:17,540
And he said it was as if I was a farmer and all the words were buried beneath the soil.

685
00:42:17,540 --> 00:42:21,500
I was constantly trying to find that one specific word in the field that contained all these different

686
00:42:21,500 --> 00:42:24,140 
words, except I didn't know where that word was.

687
00:42:24,140 --> 00:42:28,580
And so, I kept on digging and digging, just trying to locate that one word to no avail.

688
00:42:28,580 --> 00:42:37,100
So yeah, a, just like beautifully written, um, and b, I think sometimes people would ask

689
00:42:37,100 --> 00:42:41,300
me how do you know this wasn't just a straight up cognitive, like he just didn't, wasn't thinking

690
00:42:41,300 --> 00:42:42,860 
anything, didn't have anything to say.

691
00:42:42,860 --> 00:42:47,420
And it's just very apparent when you read the supplementary materials that there was lots

692
00:42:47,420 --> 00:42:51,100 
and lots and lots of very active and insightful thought happening.

693
00:42:51,100 --> 00:42:54,500
And it was just a matter of kind of bringing it to the level where it could be expressed

694
00:42:54,500 --> 00:42:56,780 
in words that was failing.

695
00:42:56,780 --> 00:43:00,260 
Um, so I think that really stuck out to me.

696
00:43:00,260 --> 00:43:05,100
Uh, I also think there's a moment where someone says he was struggling with lexical retrieval,

697
00:43:05,100 --> 00:43:09,260
but not with the word lexical retrieval, not with retrieving the brain retrieval.

698
00:43:09,260 --> 00:43:10,620 
Um, you know, he's lots of linguists.

699
00:43:10,620 --> 00:43:12,100 
That was one of his linguist friends, right?

700
00:43:12,100 --> 00:43:13,100 
Yeah, exactly.

701
00:43:13,100 --> 00:43:17,380 
Um, so I think that's, uh, a, just kind of key.

702
00:43:17,380 --> 00:43:24,300
And b, uh, gets at this, um, you know, interesting dissociation between like, what is it like

703
00:43:24,300 --> 00:43:28,940
to try to retrieve words in context versus to just have specialized language that for whatever

704
00:43:28,940 --> 00:43:31,940
reason is like at the tip of your tongue all the time, especially when you're around your

705
00:43:31,940 --> 00:43:33,620 
linguist friends, right?

706
00:43:33,620 --> 00:43:40,780
Um, and then another thing that stuck out to me actually was, um, he says something about

707
00:43:40,780 --> 00:43:47,500
appreciating when other people would fill in silences with stories or with, uh, attempts to kind

708
00:43:47,500 --> 00:43:53,100
of help them come up with the words he was thinking of, um, which I know at times, at least in

709
00:43:53,100 --> 00:43:57,900
sort of more therapeutic contexts, you're instructed to just give people time and like, not

710
00:43:57,900 --> 00:44:02,020
jump in and let them get to their own, you know, get to the words when they come.

711
00:44:02,020 --> 00:44:05,980
Um, and I think the fact that this was transient in his case might have a lot to do with this

712
00:44:05,980 --> 00:44:11,540
because he wasn't necessarily, you know, living for years and years with people jumping in all

713
00:44:11,540 --> 00:44:12,540 
the time.

714
00:44:12,540 --> 00:44:16,940
Um, but I did think that was very interesting from sort of an interpersonal perspective that,

715
00:44:16,940 --> 00:44:21,820 
uh, helping someone out is okay, um, if they're, yeah.

716
00:44:21,820 --> 00:44:27,020
I think, I think that there's, I think there's a really, so situationally dependent, like

717
00:44:27,020 --> 00:44:31,500
when to jump in and when to give space, um, when talking to people with aphasia.

718
00:44:31,500 --> 00:44:36,780
And I think that a, like a great, like a great, like a great communicator with people with

719
00:44:36,780 --> 00:44:41,980
aphasia gets this real sense of like exactly how to pull that, when to pull back and when

720
00:44:41,980 --> 00:44:47,020
to go forward with that, you know, I think that's, yeah, just something that comes with time.

721
00:44:47,020 --> 00:44:49,020 
Yeah, yeah.

722
00:44:49,020 --> 00:44:50,020 
Yeah.

723
00:44:50,020 --> 00:44:59,700
Um, okay, so you, you kind of saw the, this very unique, aphasia, you, you realized had

724
00:44:59,700 --> 00:45:04,460
that aha moment where you connected it to another individual that you'd seen and looked at her

725
00:45:04,460 --> 00:45:07,740 
scan and found that they had the same brain area.

726
00:45:07,740 --> 00:45:12,780
And so, you probably thought at that point, oh, I could, I could, I should write a paper.

727
00:45:12,780 --> 00:45:18,180
Um, and, you know, being a cognitive neuroscientist, you didn't just stick to anecdotes, um, and case

728
00:45:18,180 --> 00:45:20,540 
studies, um, you did an analysis.

729
00:45:20,540 --> 00:45:28,540
So can you tell us, um, how you decided to quantify, um, this disproportionate impairment

730
00:45:28,540 --> 00:45:31,940 
of spontaneous speech that was the hallmark of these cases?

731
00:45:31,940 --> 00:45:32,940 
Yeah.

732
00:45:32,940 --> 00:45:39,980
So we, we used, so we were using two different evaluations at different times in the history

733
00:45:39,980 --> 00:45:40,980 
of this lab.

734
00:45:40,980 --> 00:45:45,860
So like I said, these types of data were collected for, um, I believe it's like almost 15 years

735
00:45:45,860 --> 00:45:51,980
at this point that this, uh, pre-immediate post and, uh, one month post data set has been

736
00:45:51,980 --> 00:45:52,980 
acquired.

737
00:45:52,980 --> 00:46:01,580
Um, the self-generated speech deficit felt like it was most captured, in connected speech

738
00:46:01,580 --> 00:46:04,740 
because that's kind of the only time you can observe it.

739
00:46:04,740 --> 00:46:11,300
Um, so for the, the lab that scoring maps on to fluency, basically, there's a fluency

740
00:46:11,300 --> 00:46:19,340
rating that's kind of clinician, um, determined for when I asked spontaneous questions, um, that

741
00:46:19,340 --> 00:46:22,620 
require spontaneous speeches and answer.

742
00:46:22,620 --> 00:46:27,740
You sort of rate from, you know, zero to 10, like how, how fluent did that sound?

743
00:46:27,740 --> 00:46:32,100
And there are, you know, thoughts of thoughts about whether that is an ordinal scale or, you

744
00:46:32,100 --> 00:46:33,100 
know, um, yeah.

745
00:46:33,100 --> 00:46:34,100 
Okay.

746
00:46:34,100 --> 00:46:35,100 
So, yeah.

747
00:46:35,100 --> 00:46:36,100 
So, yeah.

748
00:46:36,100 --> 00:46:39,580
So the earlier patients were tested on the web or Western or phasor battery and for them,

749
00:46:39,580 --> 00:46:46,620
you use the fluency scale, which is a truly horrible scale that I really love like, I

750
00:46:46,620 --> 00:46:49,860 
love teaching it to students and tearing it apart and like help.

751
00:46:49,860 --> 00:46:54,140
I think that like, when I teach students about all the reasons why the web fluency scale is

752
00:46:54,140 --> 00:46:58,980
bad, it, it really gives them insights about assessment and like what an assessment should

753
00:46:58,980 --> 00:47:01,180 
be and shouldn't be, but all, but I'll not say it.

754
00:47:01,180 --> 00:47:03,220 
I understand why you chose it from your data.

755
00:47:03,220 --> 00:47:04,500 
That's what you had.

756
00:47:04,500 --> 00:47:06,900 
Um, and it doesn't make sense.

757
00:47:06,900 --> 00:47:07,900 
Yeah.

758
00:47:07,900 --> 00:47:08,900 
Right.

759
00:47:08,900 --> 00:47:09,900 
Yeah.

760
00:47:09,900 --> 00:47:11,900
And then at some point, I convinced Eddie to switch over to the quick phasor battery.

761
00:47:11,900 --> 00:47:14,020 
So the second half of your data set would be that.

762
00:47:14,020 --> 00:47:15,940 
So what did you do with those patients?

763
00:47:15,940 --> 00:47:19,940
Yeah.

764
00:47:19,940 --> 00:47:25,620
So we used the reduced, reduced rate, reduced length and overall communication impairment,

765
00:47:25,620 --> 00:47:30,820
ratings from the spontaneous speech, connected speech part of the quick phasor battery because

766
00:47:30,820 --> 00:47:38,260
it kind of captured the elements that we were most interested in studying from the perspective

767
00:47:38,260 --> 00:47:44,180
of like spontaneous speech that is, you know, sounds unnaturally slow and unnaturally kind

768
00:47:44,180 --> 00:47:49,540
of without content, basically without just there isn't much of it, the sparseness of the

769
00:47:49,540 --> 00:47:51,540 
spontaneous speech.

770
00:47:51,540 --> 00:47:58,100
So we took the average of those three ratings and used those as something to map this specific

771
00:47:58,100 --> 00:48:04,100
issue generating speech spontaneously and independently as captured through the connected

772
00:48:04,100 --> 00:48:05,100 
speech questions.

773
00:48:05,100 --> 00:48:06,100 
Cool.

774
00:48:06,100 --> 00:48:07,100 
Yeah.

775
00:48:07,100 --> 00:48:08,100 
So, yeah.

776
00:48:08,100 --> 00:48:12,140
So, you've kind of basically with the two different batteries you've in each case got a mechanism

777
00:48:12,140 --> 00:48:18,620
for deriving a scalar number that's that quantifies their, you know, difficulty generating

778
00:48:18,620 --> 00:48:21,940 
spontaneous speech.

779
00:48:21,940 --> 00:48:25,140
And yeah, like it makes, it all, it makes plenty of sense to me how you did it with the

780
00:48:25,140 --> 00:48:27,140 
data, the data set that you had.

781
00:48:27,140 --> 00:48:28,620 
I think that's always a challenge, right?

782
00:48:28,620 --> 00:48:31,860
It's like, you know, we've got this inside this thing that we want to describe and it's

783
00:48:31,860 --> 00:48:33,100 
like, well, how do we quantify it?

784
00:48:33,100 --> 00:48:35,700 
So that's how you quantified it.

785
00:48:35,700 --> 00:48:38,260 

And then you related it to their lesions. 786
00:48:38,260 --> 00:48:41,140 
So can you talk about how you did that?

787
00:48:41,140 --> 00:48:42,140 
Yeah.

788
00:48:42,140 --> 00:48:46,580
So this was with a voxel based lesion symptom mapping, which you are intimately familiar

789
00:48:46,580 --> 00:48:48,260 
with.

790
00:48:48,260 --> 00:48:56,820
And so, this essentially takes the integrity of a given voxel in the brain and checks if

791
00:48:56,820 --> 00:49:01,260
there is a difference between the behavior of people who do or don't have that voxel

792
00:49:01,260 --> 00:49:03,740 
as part of their lesion.

793
00:49:03,740 --> 00:49:06,900 
So it does that across all of the voxels in the brain.

794
00:49:06,900 --> 00:49:13,640
And then what results is a map of the voxels where those differences are pronounced, the

795
00:49:13,640 --> 00:49:17,100 
differences in a behavior given whether or not there's a lesion there.

796
00:49:17,100 --> 00:49:23,460
So, in our case, we use that spontaneous speech measure, the self-
generated speech measure

797
00:49:23,460 --> 00:49:28,180 
that comes from the WAB and the QAB as our behavior of interest.

798
00:49:28,180 --> 00:49:35,460 
And then we included overall QAB score or AQ as the, as a covariate.

799
00:49:35,460 --> 00:49:41,820
So, we can kind of account for the fact that this thing is going to be low while the overall

800
00:49:41,820 --> 00:49:44,180 
score is high.

801
00:49:44,180 --> 00:49:50,020
And we also co-variate out whether a proxy of speech was present or not because the presentation

802
00:49:50,020 --> 00:49:52,780
we were interested in was specifically not a motor speech presentation.

803
00:49:52,780 --> 00:49:57,460 
Like that was not the reason behind the diminished output.

804
00:49:57,460 --> 00:50:06,340
So yeah, we did VLSM and we found this really clean pre-SMA region of interest that's in

805
00:50:06,340 --> 00:50:08,780 
the figure, figure two of the paper.

806
00:50:08,780 --> 00:50:10,780 
Yeah, figure two A.

807
00:50:10,780 --> 00:50:11,780 
Yeah, figure two A.

808
00:50:11,780 --> 00:50:12,780
It's very clean.

809
00:50:12,780 --> 00:50:13,780 
It's very nice.

810
00:50:13,780 --> 00:50:16,940 
Yeah, it's super, super clean.

811
00:50:16,940 --> 00:50:21,420
And then we did the same thing with multivariate lesion symptom mapping, which is this kind of

812
00:50:21,420 --> 00:50:27,140
newer version of this type of analysis that instead of treating each voxel independently,

813
00:50:27,140 --> 00:50:32,380
it kind of uses them all as part of one big model and back projects onto each voxel like

814
00:50:32,380 --> 00:50:36,860 
the weight that it seems to have on the outcome of the behavior.

815
00:50:36,860 --> 00:50:39,740 
And they're very, very similar in their findings.

816
00:50:39,740 --> 00:50:40,740 
So that is.

817
00:50:40,740 --> 00:50:46,660
Yeah, as has been the case in I think every MLSM study that's been done.

818
00:50:46,660 --> 00:50:50,980
And I think it's kind of mysterious as to why it's so similar when like conceptually

819
00:50:50,980 --> 00:50:53,780 
it seems so much better like MLSM.

820
00:50:53,780 --> 00:50:55,820 
Yeah, but it just always gets the same result.

821
00:50:55,820 --> 00:51:02,140
And you obviously had used MLSM in your dissertation as well, so it made sense that you.

822
00:51:02,140 --> 00:51:04,380 
Yeah, although for a totally different purpose.

823
00:51:04,380 --> 00:51:08,540 
Yeah, it was about the outcome not about mapping the lesion base.

824
00:51:08,540 --> 00:51:10,740 
True, true.

825
00:51:10,740 --> 00:51:17,380
And maybe MLSM is actually really useful for outcome prediction, whereas for like leisure

826
00:51:17,380 --> 00:51:22,100
localize it, like behavior, like lesions into mapping, it actually kind of tends to just

827
00:51:22,100 --> 00:51:25,620 
give the same results you might get from VLSM.

828
00:51:25,620 --> 00:51:29,500
Yeah, I think there might be differences between whether there's like a particular brain

829
00:51:29,500 --> 00:51:32,860 
area that you suspect is involved versus a network of brain areas.

830
00:51:32,860 --> 00:51:37,260
Sometimes MLSM might be better for that, but in principle, but has anybody ever shown

831
00:51:37,260 --> 00:51:38,260 
it?

832
00:51:38,260 --> 00:51:43,700
Ivanova has a paper where she kind of talks about it, but I think she is of

833
00:51:43,700 --> 00:51:46,860 
the same opinion as us, like largely it's going to be the same.

834
00:51:46,860 --> 00:51:47,860 
Yeah.

835
00:51:47,860 --> 00:51:49,740 
Okay, so that's nice.

836
00:51:49,740 --> 00:51:53,100
So yeah, it replicates within MSLM, but and it's super clean.

837
00:51:53,100 --> 00:51:58,260
And then you have this one last analysis where you look at the relative risk kind of this

838
00:51:58,260 --> 00:52:03,980
sort of almost like a chi square type analysis of like having this pre-SMA damage and having

839
00:52:03,980 --> 00:52:06,940 
this behavioral manifestation, this unique kind of aphasia.

840
00:52:06,940 --> 00:52:09,580 
So what do you see there in your data set?

841
00:52:09,580 --> 00:52:16,700
Yeah, so basically, we looked at people who did have the resection and either did or didn't

842
00:52:16,700 --> 00:52:22,900
have the spontaneous speech deficit or people who did have the spontaneous speech deficit.

843
00:52:22,900 --> 00:52:24,500 
And either did or didn't have the resection.

844
00:52:24,500 --> 00:52:26,300 
Those were kind of the conditions of interest.

845
00:52:26,300 --> 00:52:35,020
So yeah, we find that you're basically if you have this deficit, you're 15 times more

846
00:52:35,020 --> 00:52:42,460
likely to have had this resection, that's a very kind of again, clear result.

847
00:52:42,460 --> 00:52:47,260
And I think what's kind of interesting too is if you look at figure three A versus figure

848
00:52:47,260 --> 00:52:51,940
three B, most of the people who had this resection and didn't have this deficit just kind

849
00:52:51,940 --> 00:52:53,220 
of didn't have a deficit.

850
00:52:53,220 --> 00:52:56,020 
Like either you're going to have nothing or you're going to have this.

851
00:52:56,020 --> 00:52:57,020 
Okay.

852
00:52:57,020 --> 00:52:58,020 
Okay.

853
00:52:58,020 --> 00:53:04,700
And so, about half the people with the pre-SMA resection to have the syndrome you describe,

854
00:53:04,700 --> 00:53:06,100 
are we calling it dynamic aphasia?

855
00:53:06,100 --> 00:53:09,980 
Like what's your preferred name for it when you think about it now?

856
00:53:09,980 --> 00:53:11,580 
Yeah, it's a great question.

857
00:53:11,580 --> 00:53:14,700 
I think I would be comfortable calling it dynamic aphasia.

858
00:53:14,700 --> 00:53:19,340
I was very careful in the paper to try not to be too tied to any particular tradition

859
00:53:19,340 --> 00:53:23,180
of thought around it, just because I didn't want to step on any toes where, you know, there's

860
00:53:23,180 --> 00:53:27,820
different theoretical assumptions, but I think dynamic aphasia is pretty clearly like the

861
00:53:27,820 --> 00:53:30,140 
clearest map onto what we observed.

862
00:53:30,140 --> 00:53:31,140 
Okay.

863
00:53:31,140 --> 00:53:36,500
So, half of them had dynamic aphasia and you're saying the other half had nothing more or less.

864
00:53:36,500 --> 00:53:39,940 
Or had sort of dynamic aphasia plus a motor speech deficit.

865
00:53:39,940 --> 00:53:40,940 
Oh, okay.

866
00:53:40,940 --> 00:53:41,940 
Yeah, okay.

867
00:53:41,940 --> 00:53:47,660
Yeah, because you required no, apraxia of speech to meet your core diagnostic criteria.

868
00:53:47,660 --> 00:53:52,020
And then you also saw that you occasionally saw dynamic aphasia in people with lesions

869
00:53:52,020 --> 00:53:53,420 
other than the SMA, right?

870
00:53:53,420 --> 00:53:55,980 
So what, who were those people?

871
00:53:55,980 --> 00:53:56,980 
Yeah.

872
00:53:56,980 --> 00:54:02,620
So, there's a sort of small trend towards it maybe being inferior frontal gyrus, but that's

873
00:54:02,620 --> 00:54:05,740 
a much smaller number of individuals who presented with this.

874
00:54:05,740 --> 00:54:11,820
So, you can see that on the color bar between A and C basically that in panel A of that figure,

875
00:54:11,820 --> 00:54:12,820 
there's three.

876
00:54:12,820 --> 00:54:13,820 
Yeah.

877
00:54:13,820 --> 00:54:14,820 
Figure three.

878
00:54:14,820 --> 00:54:15,820 
Okay.

879
00:54:15,820 --> 00:54:21,700
There's a very clear kind of hotspot in this pre-SMA area where, you know, online people

880
00:54:21,700 --> 00:54:25,420 
that have that perception and also have that deficit fall right there.

881
00:54:25,420 --> 00:54:31,220
Whereas in figure 3C, you have about three people who have this deficit where it kind

882
00:54:31,220 --> 00:54:35,620
of centers on the inferior frontal gyrus, which might be meaningful, but it's certainly

883
00:54:35,620 --> 00:54:38,900 
not everybody else with the deficit has an inferior frontal lesion.

884
00:54:38,900 --> 00:54:39,900 
Yeah.

885
00:54:39,900 --> 00:54:41,500 
It can be a little more widespread than that.

886
00:54:41,500 --> 00:54:42,500 
Okay.

887
00:54:42,500 --> 00:54:47,460
But it's really like pre-SMA is really the region that is much more strongly associated

888
00:54:47,460 --> 00:54:49,460 
with this than anything else.

889
00:54:49,460 --> 00:54:53,220 
At least based on the way we did, yeah.

890
00:54:53,220 --> 00:54:54,220 
Yeah.

891
00:54:54,220 --> 00:54:57,780 
And the surgical population is kind of good for this question, right?

892
00:54:57,780 --> 00:55:03,580
I mean, they offer you various advantages relative to other neurological conditions.

893
00:55:03,580 --> 00:55:04,580 
Yeah.

894
00:55:04,580 --> 00:55:05,580 
Yeah.

895
00:55:05,580 --> 00:55:08,580 
I mean, I think there's the precision of the lesions.

896
00:55:08,580 --> 00:55:13,620
I mean, it's not going to be just kind of a natural experiment based on, you know, where

897
00:55:13,620 --> 00:55:17,940 
the, an occlusion occurs in an artery or something like that.

898
00:55:17,940 --> 00:55:20,580 
And I think ACA occlusions are actually quite rare.

899
00:55:20,580 --> 00:55:21,580
Yeah.

900
00:55:21,580 --> 00:55:23,060 
They are quite rare.

901
00:55:23,060 --> 00:55:24,060 
Yeah.

902
00:55:24,060 --> 00:55:26,340 
So you don't have as many opportunities to study it in stroke.

903
00:55:26,340 --> 00:55:30,300
And when you do, it's rare that it's going to only damage this area because it's going

904
00:55:30,300 --> 00:55:33,060 
to be kind of depending on where the occlusion is.

905
00:55:33,060 --> 00:55:37,860
It can also affect the SMA, it can affect, you know, frontal areas in front of it, or,

906
00:55:37,860 --> 00:55:41,060 
you know, laterally, so, probably more so laterally.

907
00:55:41,060 --> 00:55:42,380 
But, yeah.

908
00:55:42,380 --> 00:55:48,980
So first of all, it gives you kind of opportunities to see precise lesions in that area.

909
00:55:48,980 --> 00:55:53,220
It also gives you an opportunity to do preoperative evaluation, which you generally are not going

910
00:55:53,220 --> 00:55:57,120
to have in stroke because you're not going to see people before they've had a stroke and

911
00:55:57,120 --> 00:55:58,120 
evaluate their language.

912
00:55:58,120 --> 00:56:00,300 
There's usually no reason for that to occur.

913
00:56:00,300 --> 00:56:07,740
So, it offers you that as well as the opportunity to interview people like right after they have had a surgery where there hasn’t been any reorganization 

914
00:56:12,460 --> 00:56:16,020 
And a month later, when often, it's their return to baseline.

915
00:56:16,020 --> 00:56:21,900 
So you get this kind of whole trajectory, which is really interesting.

916
00:56:21,900 --> 00:56:29,100
And neurodegenerative populations, neurodegenerative populations, you can also see some patterns

917
00:56:29,100 --> 00:56:30,300 
sort of like this.

918
00:56:30,300 --> 00:56:34,140 
But again, it's going to rarely target just that one area.

919
00:56:34,140 --> 00:56:39,620
And it's, you're not going to get these opportunities to hear about recovery as well, unfortunately,

920
00:56:39,620 --> 00:56:41,740 
because so, yeah.

921
00:56:41,740 --> 00:56:46,220
So, it is a really interesting and unique population to get to learn about this from.

922
00:56:46,220 --> 00:56:47,220
Okay.

923
00:56:47,220 --> 00:56:48,220 
Yeah.

924
00:56:48,220 --> 00:56:50,660 
Yeah, these transient aphasias are kind of really focal.

925
00:56:50,660 --> 00:56:52,420 
That's like their value.

926
00:56:52,420 --> 00:56:54,100 
But also a mystery.

927
00:56:54,100 --> 00:56:56,620 
But we'll talk about that in a second.

928
00:56:56,620 --> 00:57:03,580
Before that, like, you know, so, yeah, pre-SMA damage causes dynamic aphasia, to oversimplify

929
00:57:03,580 --> 00:57:04,580 
maybe.

930
00:57:04,580 --> 00:57:07,940 
Would you say that pre-SMA is therefore a language region?

931
00:57:07,940 --> 00:57:10,820 
Like, how do you end up coming down on that point?

932
00:57:10,820 --> 00:57:14,140 
Yeah, it's such a good question.

933
00:57:14,140 --> 00:57:19,380
I think that it's, and other people have said this before, I think it's really at that boundary

934
00:57:19,380 --> 00:57:20,900
between language and thought.

935
00:57:20,900 --> 00:57:28,700
I think there's been a lot of work suggesting that it's involved in either kind of like pushing

936
00:57:28,700 --> 00:57:30,780 
choices to the surface.

937
00:57:30,780 --> 00:57:34,660 
This has shown up in animal work.

938
00:57:34,660 --> 00:57:37,260 
It's shown up in the fluency tasks that are done by the Robinson lab.

939
00:57:37,260 --> 00:57:44,580
It seems to be possibly domain generally, just any time you have kind of like a wide space

940
00:57:44,580 --> 00:57:47,100 
of possibilities where you can do anything.

941
00:57:47,100 --> 00:57:53,140
It seems like it might be involved in sort of increasing the activation of any given arbitrary

942
00:57:53,140 --> 00:57:57,100 
choice within like a field of possible choices.

943
00:57:57,100 --> 00:58:01,860
And when that happens in sort of the language system, I think the idea is that you have this

944
00:58:01,860 --> 00:58:06,100
linguistic apparatus that's functioning fine, but if you don't have any specific cue

945
00:58:06,100 --> 00:58:11,900
or input about like what's worth discussing, like what rises above the threshold of, you

946
00:58:11,900 --> 00:58:17,540
know, like this is relevant to say right now, there's possibly just nothing that's pushing

947
00:58:17,540 --> 00:58:21,140 
anything to that level of like that is the thing that you should say.

948
00:58:21,140 --> 00:58:23,740 
That is the thing that is worth expressing right now.

949
00:58:23,740 --> 00:58:29,140 
Is that Gail Robinson's energization concept?

950
00:58:29,140 --> 00:58:35,180
I'm kind of using I think energization and also response selection or like task monitoring.

951
00:58:35,180 --> 00:58:38,180
And this is something at some point I would love to talk with her about like, you know,

952
00:58:38,180 --> 00:58:45,660
what is the kind of like really clear distinction between those things behaviorally because

953
00:58:45,660 --> 00:58:52,180
I think energization is like initiating the response and sustaining it over time, which

954
00:58:52,180 --> 00:58:54,100 
I think you can describe in a similar way, right?

955
00:58:54,100 --> 00:58:57,100
Like you can say first you have to decide something is worth saying and then you have to

956
00:58:57,100 --> 00:58:59,260
keep deciding to continue with it.

957
00:58:59,260 --> 00:59:06,340
Like there is this kind of continuous role of some process in deciding like, this is

958
00:59:06,340 --> 00:59:10,460
the thing to say now that you've said that this is the next thing to say you should keep

959
00:59:10,460 --> 00:59:14,980
going, you know, there's this sort of like decision making process around that process

960
00:59:14,980 --> 00:59:17,300 
moving forward.

961
00:59:17,300 --> 00:59:22,980
Whereas response selection or the task monitoring stuff is more about when there are higher

962
00:59:22,980 --> 00:59:30,220
low constraints around what you could say, like selecting between competing options, which,

963
00:59:30,220 --> 00:59:34,620
you know, I can see, I can see those kind of being based on a similar mechanism.

964
00:59:34,620 --> 00:59:37,780 
I can see them playing out in different ways.

965
00:59:37,780 --> 00:59:42,780
But I think the energization thing maybe crucially is more domain general according to this perspective.

966
00:59:42,780 --> 00:59:49,940
Like any kind of task whether it's verbal or gesture or drawing pictures like all of those

967
00:59:49,940 --> 00:59:55,260
would be affected by a Pre-SMA lesion or a medial frontal lesion in her theories.

968
00:59:55,260 --> 01:00:00,660
Whereas the sort of like high low constraint differences when there's, you know, maybe like

969
01:00:00,660 --> 01:00:06,100
higher closed probability for a given sense or something, those would be more associated

970
01:00:06,100 --> 01:00:10,340 
with lateral frontal regions and would be more of language specific.

971
01:00:10,340 --> 01:00:16,140
So, yeah, so the question of like, is this a language impairment or is this part of the

972
01:00:16,140 --> 01:00:17,140 
language system?

973
01:00:17,140 --> 01:00:23,220
I think when it comes to natural language and like producing discourse that is functional

974
01:00:23,220 --> 01:00:25,580 
in the world, you need this region.

975
01:00:25,580 --> 01:00:33,420 
I think that that I don't feel hesitant about saying it all.

976
01:00:33,420 --> 01:00:37,500
Whether this should be considered part of like the mechanics that support language, I

977
01:00:37,500 --> 01:00:39,580 
think this is something that sort of interfaces with that.

978
01:00:39,580 --> 01:00:46,140
I think this is something that pulls into that language system and pulls linguistic

979
01:00:46,140 --> 01:00:49,820
constructs to the surface to be expressed or, you know, kind of interfaces between the

980
01:00:49,820 --> 01:00:52,860 
thoughts themselves and that language system.

981
01:00:52,860 --> 01:00:53,860 
Do you have thoughts?

982
01:00:53,860 --> 01:00:55,820 
No, I think I agree with you.

983
01:00:55,820 --> 01:00:56,820 
Yeah, I think.

984
01:00:56,820 --> 01:00:59,020 
I guess so, yeah, I have thoughts.

985
01:00:59,020 --> 01:01:04,180
Like, I mean, one thing about the domain generality of a, like one thing that's really striking

986
01:01:04,180 --> 01:01:07,540 
is so you only have left hemisphere patients in your cohort.

987
01:01:07,540 --> 01:01:13,260
But in Binder's meta-analysis, it's super, super-lateralized, like the role of this region

988
01:01:13,260 --> 01:01:15,300 
in semantics.

989
01:01:15,300 --> 01:01:20,740
So, you know, to the extent that it's, you know, it probably is only the left that's relevant

990
01:01:20,740 --> 01:01:21,740 
for language.

991
01:01:21,740 --> 01:01:27,180 
So there probably is at least some specialization of its role.

992
01:01:27,180 --> 01:01:33,860
And I guess, well, you know, if you take seriously Edwin's description of the search for the

993
01:01:33,860 --> 01:01:39,540
items, like buried in like a farmer with buried in the field, like it's, it's not really a,

994
01:01:39,540 --> 01:01:43,540 
it's definitely not a sort of choosing among available choices, right?

995
01:01:43,540 --> 01:01:47,700
It's like, it's like finding anything that's, that's like, yeah, exactly.

996
01:01:47,700 --> 01:01:51,260
And you sort of specifically, I actually asked him at one point, we met with him about six

997
01:01:51,260 --> 01:01:55,340
months afterwards and asked him, like, do you remember having trouble selecting between options

998
01:01:55,340 --> 01:01:56,340 
on the menu?

999
01:01:56,340 --> 01:02:00,100 
Do you remember, you know, struggling to make the, any didn't?

1000
01:02:00,100 --> 01:02:03,500
And he actually specifically, I think I later sent him a draft of some lab meeting

1001
01:02:03,500 --> 01:02:09,220
slides I had where I was going to say, like, you know, maybe, maybe even if it doesn't feel

1002
01:02:09,220 --> 01:02:11,780 
like it's a selection deficit, it's still a selection deficit.

1003
01:02:11,780 --> 01:02:13,820 
And he was like, it's not a select deficit.

1004
01:02:13,820 --> 01:02:16,220 
It was really confident that it wasn't.

1005
01:02:16,220 --> 01:02:18,060 
That's so really cool.

1006
01:02:18,060 --> 01:02:19,060 
Yeah.

1007
01:02:19,060 --> 01:02:24,700
But again, I mean, there is, I think the, the personal experience versus the like theoretical

1008
01:02:24,700 --> 01:02:28,940
possibility, it's, you know, it's hard to disentangle, you know, what somebody has access to in their

1009
01:02:28,940 --> 01:02:31,700 
own lexicon and search.

1010
01:02:31,700 --> 01:02:34,980 
But, you know, I'm inclined to trust him on that.

1011
01:02:34,980 --> 01:02:41,420
Yeah, but I think with that searching in the soil analogy, like, there's, what I'm picturing

1012
01:02:41,420 --> 01:02:45,140 
is kind of like, this is a messy metaphor.

1013
01:02:45,140 --> 01:02:47,740 
You can decide whether to keep this in or not.

1014
01:02:47,740 --> 01:02:51,820
But if you imagine under the soil where he's trying to pick the, the plants or the crops,

1015
01:02:51,820 --> 01:02:55,660
that there's something that normally pushes some of those crops closer to the surface,

1016
01:02:55,660 --> 01:02:58,020 
where it's like, that's the one that's relevant here.

1017
01:02:58,020 --> 01:03:04,340
And it seems like that, that sort of elevator underneath the soil just wasn't there anymore.

1018
01:03:04,340 --> 01:03:10,780
So, yeah, it's not that there was a, it's not that there weren't words to be found is

1019
01:03:10,780 --> 01:03:14,780
that those words weren't being pushed to the surface to be selected for expression.

1020
01:03:14,780 --> 01:03:15,780 
Definitely keeping that in.

1021
01:03:15,780 --> 01:03:17,260 
I think that's a great metaphor.

1022
01:03:17,260 --> 01:03:20,380 
Yeah, hopefully they'll forgive the elevator under the soil.

1023
01:03:20,380 --> 01:03:26,100 
That's, you know, we, you know, it's a conversation.

1024
01:03:26,100 --> 01:03:27,700 
You came up with it on the fly.

1025
01:03:27,700 --> 01:03:34,940
So, the aphasias that we see in these post surgical patients are transient usually,

1026
01:03:34,940 --> 01:03:40,700
they're usually largely resolved by a month with just a few residual issues as you discuss

1027
01:03:40,700 --> 01:03:43,300 
and we, has been shown in other work.

1028
01:03:43,300 --> 01:03:48,300 
Isn't it interesting that something can cause such a profound deficit?

1029
01:03:48,300 --> 01:03:53,860 
And yet, the brain is able to find out another way round like this.

1030
01:03:53,860 --> 01:03:56,860 
And we don't really know where that other way round is, right?

1031
01:03:56,860 --> 01:04:04,420
Yeah, it'd be a really interesting fMRI study, I guess, to look at sort of within a month

1032
01:04:04,420 --> 01:04:09,660
or so, how, how is this area getting re reintegrated and, or, you know, this, this function

1033
01:04:09,660 --> 01:04:11,820 
getting reintegrated through other areas?

1034
01:04:11,820 --> 01:04:18,700
Yeah, when, and this, and in our, the task that we use that you've worked with a lot,

1035
01:04:18,700 --> 01:04:24,940
it does activate the, this area, it does activate the medial surface of the frontal lobe.

1036
01:04:24,940 --> 01:04:29,700 
So, yeah, adaptive language mapping, yeah.

1037
01:04:29,700 --> 01:04:35,340
So, you know, if we scan somebody like Edwin, when we would not see that, we would not see,

1038
01:04:35,340 --> 01:04:39,740 
we'd see at least a whole where they should have been in activation.

1039
01:04:39,740 --> 01:04:46,420
Whether we would see residual activation around his resection, we might, or would we just

1040
01:04:46,420 --> 01:04:50,740
see nothing, would we just see like the rest of the language network activating as normal

1041
01:04:50,740 --> 01:04:56,820
and the, you know, deficit has been overcome and we don't understand how that happened.

1042
01:04:56,820 --> 01:04:58,980 
That's probably, that's what I'm gonna...

1043
01:04:58,980 --> 01:05:01,380 
Yeah, we can get a surprise right hemisphere.

1044
01:05:01,380 --> 01:05:06,180
Yeah, but that would not be my expectation, but that would be the most exciting finding,

1045
01:05:06,180 --> 01:05:07,180 
yeah, definitely.

1046
01:05:07,180 --> 01:05:10,140
That would be, that would be by far the most exciting, but it's not something that we frequently

1047
01:05:10,140 --> 01:05:11,140 
seen.

1048
01:05:11,140 --> 01:05:12,140 
Yeah.

1049
01:05:12,140 --> 01:05:15,340 
Yeah, so it's kind of mysterious, right?

1050
01:05:15,340 --> 01:05:19,260 
Like, how do these individuals recover?

1051
01:05:19,260 --> 01:05:25,020 
Yeah, I think that's still an open question.

1052
01:05:25,020 --> 01:05:27,460 
Yeah, definitely.

1053
01:05:27,460 --> 01:05:29,580 
Last thing about the paper.

1054
01:05:29,580 --> 01:05:36,140
So, it starts with the line, "For an aphasia-friendly version of this article, please see," and then

1055
01:05:36,140 --> 01:05:41,980
you have a link to a, to a version of the paper, which is written in a way that's accessible

1056
01:05:41,980 --> 01:05:42,980 
to people with aphasia.

1057
01:05:42,980 --> 01:05:44,220 
It has simple language.

1058
01:05:44,220 --> 01:05:47,020 
It has iconography.

1059
01:05:47,020 --> 01:05:53,860
Can you tell us about why you make aphasia-friendly versions of all your papers and how you go about

1060
01:05:53,860 --> 01:05:55,860 
it and why you think it's important?

1061
01:05:55,860 --> 01:05:56,860 
Yeah, yeah.

1062
01:05:56,860 --> 01:06:00,380
So, I touched on this a little bit earlier with the aphasia group stuff.

1063
01:06:00,380 --> 01:06:06,700
I think that there's so much curiosity about what people are learning about aphasia, among

1064
01:06:06,700 --> 01:06:11,660
people with aphasia, and it's so hard for them to get that information, especially at sort

1065
01:06:11,660 --> 01:06:17,140
of the researcher-generated level as opposed to press releases or coverage.

1066
01:06:17,140 --> 01:06:23,060
So, that's something that Anna Kasdan and I got really passionate about in grad school,

1067
01:06:23,060 --> 01:06:26,660 
and I've tried to carry it through with me as I continue publishing.

1068
01:06:26,660 --> 01:06:31,940
So, yeah, I mean, especially for people where you're undergoing like an elective surgery,

1069
01:06:31,940 --> 01:06:35,740
or you have experienced something after surgery, or even if you've just had a stroke and it

1070
01:06:35,740 --> 01:06:40,380 
happens to, you know, align with something that has been studied.

1071
01:06:40,380 --> 01:06:43,660 
I think really great to be able to get from the researcher's mouth.

1072
01:06:43,660 --> 01:06:46,060 
Like, here's what I think you should understand about this.

1073
01:06:46,060 --> 01:06:49,300 
Here's what might be relevant for you and your family.

1074
01:06:49,300 --> 01:06:56,380
So, that's been a big part of what I try to and hope to continue trying to do as a researcher.

1075
01:06:56,380 --> 01:07:01,420
Like, make that research not just academic and actually get it out to the people that

1076
01:07:01,420 --> 01:07:03,420 
it's about.

1077
01:07:03,420 --> 01:07:09,060
So, in terms of making them at the time of this that I was drafting this one, I basically

1078
01:07:09,060 --> 01:07:12,900
just opened up a Google doc and I try to think like, okay, what are the key messages here?

1079
01:07:12,900 --> 01:07:17,220 
And, you know, where do I find free icons that demonstrate it?

1080
01:07:17,220 --> 01:07:18,980 
And that for me is kind of a fun process.

1081
01:07:18,980 --> 01:07:21,260 
I really like doing that.

1082
01:07:21,260 --> 01:07:27,340
But actually Anna and me and my husband Isaac, who I met at the computational memory lab,

1083
01:07:27,340 --> 01:07:34,940
he's a software developer, have now started working on an LLM-based version of this.

1084
01:07:34,940 --> 01:07:42,540
And, there was a lot of the upfront work for you with the extreme caveat that a researcher

1085
01:07:42,540 --> 01:07:46,460
who uses that absolutely has to check everything that comes out of it and the icons are going

1086
01:07:46,460 --> 01:07:49,300 
to require a lot of tweaking.

1087
01:07:49,300 --> 01:07:53,180
But it's sort of a way I'm hoping to motivate people to do this type of thing because it

1088
01:07:53,180 --> 01:07:57,460
might not feel like such a big lift if there's been a touch of the work done for you as

1089
01:07:57,460 --> 01:07:58,460 
a set.

1090
01:07:58,460 --> 01:08:01,260 
So, you get an LLM to write the first draft?

1091
01:08:01,260 --> 01:08:02,260 
Yeah.

1092
01:08:02,260 --> 01:08:04,260 
And then you, who get yourself?

1093
01:08:04,260 --> 01:08:07,100 
Yeah, lovely.

1094
01:08:07,100 --> 01:08:09,340 
So yeah, this is a great paper.

1095
01:08:09,340 --> 01:08:12,140 
I think everybody should read it.

1096
01:08:12,140 --> 01:08:16,940
And it's just like a beautiful description of an aphasia syndrome that doesn't get talked

1097
01:08:16,940 --> 01:08:23,580
about that much, but is very interesting and teaches us something about the language network.

1098
01:08:23,580 --> 01:08:30,460
So, the last thing I wanted to talk about beyond the paper is your current job.

1099
01:08:30,460 --> 01:08:35,620
So you're now a lecturer for the Princeton Writing Program, which I know is a job that you really

1100
01:08:35,620 --> 01:08:36,900 
love.

1101
01:08:36,900 --> 01:08:44,020
Can you tell us about how you came into that job and what led you in that direction?

1102
01:08:44,020 --> 01:08:45,020 
Yeah.

1103
01:08:45,020 --> 01:08:49,860 
So, I always loved teaching, always, always, always.

1104
01:08:49,860 --> 01:08:55,540
And when I was at Vanderbilt, I did the bold program, which was a center for teaching

1105
01:08:55,540 --> 01:09:00,100
based program where you got to design an online module for some existing class.

1106
01:09:00,100 --> 01:09:04,780 
I did it for the language psychology class at Vanderbilt.

1107
01:09:04,780 --> 01:09:08,340 
I did a bunch of teaching trainings at UCSF, their step-up program.

1108
01:09:08,340 --> 01:09:09,980 
I always loved teaching.

1109
01:09:09,980 --> 01:09:14,660
And as much as I loved grad school, I was actually thinking I would get more TA experience,

1110
01:09:14,660 --> 01:09:20,180
loved doing the TA for our very small class of already very qualified and brilliant master

1111
01:09:20,180 --> 01:09:21,180 
students.

1112
01:09:21,180 --> 01:09:26,940
But what I loved doing in undergrad was teaching, like, it was largely teaching freshman,

1113
01:09:26,940 --> 01:09:29,380 
actually, freshman and sophomores in the inter-site classes.

1114
01:09:29,380 --> 01:09:31,060 
So, I really liked teaching.

1115
01:09:31,060 --> 01:09:34,940 
I always knew I wanted to be involved in teaching in my career.

1116
01:09:34,940 --> 01:09:35,940 
And I always liked writing.

1117
01:09:35,940 --> 01:09:40,580
I don't think I touched on this in my sort of description of what brought me into the field.

1118
01:09:40,580 --> 01:09:46,060
But one of the ways I sort of thought about filtering the world through language was through

1119
01:09:46,060 --> 01:09:47,060 
writing.

1120
01:09:47,060 --> 01:09:49,020 
And I was the editor of a literary magazine at high school.

1121
01:09:49,020 --> 01:09:54,460
And I was very kind of involved in trying to make stories out of the world.

1122
01:09:54,460 --> 01:09:59,300
So yeah, the writing passion and the teaching passion have been there throughout my whole

1123
01:09:59,300 --> 01:10:01,260 
journey.

1124
01:10:01,260 --> 01:10:06,540
When I, it was 2022 that I went to Cold Spring Harbor for their Neurobiology of Language

1125
01:10:06,540 --> 01:10:09,540 
event, I guess.

1126
01:10:09,540 --> 01:10:14,660
It was like a week-long sort of seminar for students to work with people who are in the

1127
01:10:14,660 --> 01:10:17,300 
Neurobiology of Language, kind of like giants in the field.

1128
01:10:17,300 --> 01:10:22,700 
And I met a woman there, Srishti Nayak, who's at Vanderbilt now.

1129
01:10:22,700 --> 01:10:28,020
And she, on the first day, people kind of introduced themselves and said their career trajectories.

1130
01:10:28,020 --> 01:10:33,580
She mentioned that for two years she had taught a class at Princeton about graphic novels and

1131
01:10:33,580 --> 01:10:36,100 
the brain, graphic novels and psychology.

1132
01:10:36,100 --> 01:10:38,260 
And I was like, oh my god, that's my dream job.

1133
01:10:38,260 --> 01:10:39,260 
That sounds so cool.

1134
01:10:39,260 --> 01:10:43,020 
So I talked to her after, and I was like, what was that job?

1135
01:10:43,020 --> 01:10:45,900
What, where were you doing that?

1136
01:10:45,900 --> 01:10:47,420 
How can I do that?

1137
01:10:47,420 --> 01:10:51,660
And she told me about this Princeton writing program where essentially people from all

1138
01:10:51,660 --> 01:10:56,820
different disciplines get to design a class from scratch that teaches freshmen about

1139
01:10:56,820 --> 01:10:59,500 
scholarly writing and the process of scholarly writing.

1140
01:10:59,500 --> 01:11:03,020 
So yeah, basically, I was like, that is the job I want.

1141
01:11:03,020 --> 01:11:08,180
And the first year that it was available, I wasn't able to apply by the deadline, so I emailed

1142
01:11:08,180 --> 01:11:13,060
them and I was like, heads up, I'm applying next year, please don't lose my email.

1143
01:11:13,060 --> 01:11:20,700
And then the next year I applied and they brought me out for an interview.

1144
01:11:20,700 --> 01:11:22,700 
It was everything I dreamed it would be.

1145
01:11:22,700 --> 01:11:24,620 
It's so much fun.

1146
01:11:24,620 --> 01:11:28,460
And it gives you so much creativity around like what you get to think about, what you get

1147
01:11:28,460 --> 01:11:33,980 
to force 18 year olds to think about.

1148
01:11:33,980 --> 01:11:36,740 
Yeah, and I get to do all kinds of fun things.

1149
01:11:36,740 --> 01:11:42,340
I actually just recently bought a toy called Mind Flex from 2009, which I'm using in my class.

1150
01:11:42,340 --> 01:11:44,140 
What is that?

1151
01:11:44,140 --> 01:11:45,820 
It's, you wear a headset.

1152
01:11:45,820 --> 01:11:48,380 
It has like a single contact.

1153
01:11:48,380 --> 01:11:54,580
And theoretically, it is reading your brain waves to move a ball around a little, a little

1154
01:11:54,580 --> 01:11:57,420 
obstacle course, basically.

1155
01:11:57,420 --> 01:12:00,420 
Okay, just a question.

1156
01:12:00,420 --> 01:12:02,380 
You know, very unclear.

1157
01:12:02,380 --> 01:12:04,660 
But that makes it ripe for analysis in a writing class.

1158
01:12:04,660 --> 01:12:05,660
Does it not?

1159
01:12:05,660 --> 01:12:10,420
You have lots, you know, you can discuss about the nature of advertising and the nature of,

1160
01:12:10,420 --> 01:12:13,900 
you know, neuromania, especially in the early 2000s.

1161
01:12:13,900 --> 01:12:19,620
So, you know, there's a lot of kind of just picking interesting weird artifacts and, you know,

1162
01:12:19,620 --> 01:12:21,860 
forcing students to think about them at a high level.

1163
01:12:21,860 --> 01:12:23,940 
So, yeah, it's been a lot of fun.

1164
01:12:23,940 --> 01:12:25,620 
I could talk about it forever.

1165
01:12:25,620 --> 01:12:26,780 
That is so cool.

1166
01:12:26,780 --> 01:12:28,420 
Yeah, what are you neat?

1167
01:12:28,420 --> 01:12:31,620 
Like, you know, career, you're building for yourself.

1168
01:12:31,620 --> 01:12:34,180 
I can't wait to see what you do next.

1169
01:12:34,180 --> 01:12:35,180 
Yeah, yeah.

1170
01:12:35,180 --> 01:12:36,180
Yeah.

1171
01:12:36,180 --> 01:12:40,220 
I'm always surprising myself, so.

1172
01:12:40,220 --> 01:12:46,860
Well, I have to take my daughter to her flute workshop that is happening all this week,

1173
01:12:46,860 --> 01:12:50,780 
which is why we met super early in the morning, my time.

1174
01:12:50,780 --> 01:12:53,180 
So I will go and do that.

1175
01:12:53,180 --> 01:12:57,020 
And it was lovely talking with you and, you know, catching up.

1176
01:12:57,020 --> 01:13:01,140 
And, you know, thanks for walking us through this paper today.

1177
01:13:01,140 --> 01:13:02,140 
Yeah.

1178
01:13:02,140 --> 01:13:03,140 
Thank you so much.

1179
01:13:03,140 --> 01:13:05,780 
Thanks for thinking of me and, yeah, great to see you again.

1180
01:13:05,780 --> 01:13:06,780 
Yeah, you too.

1181
01:13:06,780 --> 01:13:07,780 
All right.

1182
01:13:07,780 --> 01:13:08,780
Take care.

1183
01:13:08,780 --> 01:13:09,780 
Bye.

1184
01:13:09,780 --> 01:13:10,780 
Okay.

1185
01:13:10,780 --> 01:13:11,780 
Well, that's it for episode 34.

1186
01:13:11,780 --> 01:13:13,300 
Thanks to Deb for joining me on the podcast.

1187
01:13:13,300 --> 01:13:20,020
And I've linked Deb's paper in the show notes and on the podcast website at langneurosci.org/podcast.

1188
01:13:20,020 --> 01:13:23,340 
Thanks also to Marcia Petyt for transcribing this episode.

1189
01:13:23,340 --> 01:13:26,340 
Please do consider submitting your papers to neurobiology of language.

1190
01:13:26,340 --> 01:13:30,540 
It's open access, society supported and has a great editorial team.

1191
01:13:30,540 --> 01:13:35,820
In my experience, I've had constructive reviews, fair decisions, and speedy publications.

1192
01:13:35,820 --> 01:13:39,420
If the article processing charge is barrier for your lab, that is always something you can

1193
01:13:39,420 --> 01:13:41,100 
talk to the editors about.

1194
01:13:41,100 --> 01:13:43,740
It's not going to stop your paper from getting published.

1195
01:13:43,740 --> 01:13:44,900 
Okay, bye for now.

1196
01:13:44,900 --> 01:13:45,540 
See you next time.