Multiprobe multiple single-unit recordings and visual information - - PowerPoint PPT Presentation

Multiprobe multiple single-unit recordings and visual information processing in the primate brain Hiroshi TAMURA

TE TEO V4 V2 V1

Saturday, March 5, 2011

Stimulus-selective responses of neurons in visual cortices

TE TEO V4 V2 V1

Desimone et al 1984 Gallant et al 1996 Ito & Komatsu 2004 Tamura &Tanaka 2001 TE V4 V2 V4 TE

Saturday, March 5, 2011

Tamura &Tanaka 2001

27.8 2.0 5.0

A neuron in the inferior temporal (IT) cortex responded to an image of a chair

Saturday, March 5, 2011

How are the selective responses of IT neurons generated ? How are the neurons involved in the representation of visual images distributed in the IT cortex ?

Saturday, March 5, 2011

Functional neuronal circuitry in the inferior temporal (IT) cortex

Saturday, March 5, 2011

Multiprobe recording allows us to record simultaneously from multiple single neurons, and to perform cross-correlation analysis of spike trains

Heptode (Thomas)

Saturday, March 5, 2011

Signals recorded with Heptode

Signals from 6 of the 7 recording probes were shown

Saturday, March 5, 2011

Isolation and classification of spikes 1 2 3 4 5

Kaneko et al. 1999, 2007

Time

cell 1 cell 2 cell n :

Spike detection Spike classification Spike trains

Saturday, March 5, 2011

Recording and visual stimulus presentation system

Saturday, March 5, 2011

trail 1 trail 2 trail 3 trail 4 trail 5 cell 1 cell 2 R

5 10 15 20 25 30

trail 1 trail 2 trail 3 trail 4 trail 5 cell 1 cell 2 R

5 10 15 20 25 30

• t

+t

• t

+t

Estimate functional connectivity with cross-correlation analysis

Time (reference) (reference)

raw correlograms shift predictors

(trials were shifted for each stimulus condition)

Saturday, March 5, 2011

0.1

• 0.1

Cell 2 -> Cell 1 Cell 1 -> Cell 2 250 250

Spikes/bin Time (s)

Cell 2 -> Cell 1 Cell 1 -> Cell 2

• 0.04

0.04 200 200

Spikes/bin Time (s)

inhibitory neuron (cell 1) target neuron (cell 2) excitatory neuron (cell 3) target neuron (cell 4)

Raw correlogram Shift predictor (control)

Inhibitory and excitatory interactions between neurons in the IT cortex

Raw correlogram Shift predictor (control)

cell 1 -> cell 2 cell 2 -> cell 1 cell 3 -> cell 4 cell 4 -> cell 3

Saturday, March 5, 2011

Inhibitory interactions were found in neuron pairs with negatively correlated stimulus preferences

A

0.1

• 0.1

Cell 2 -> Cell 1 Cell 1 -> Cell 2

Spikes/s Spikes/s Stimulus

B

1 67 Cell 1 Cell 2 30

• 20

10

• 15

250 250

Spikes/bin Time (s)

฀ ฀

Signal correlation

Tamura et al. 2003, 2004

Saturday, March 5, 2011

Excitatory interactions were found in neuron pairs with positively correlated stimulus preferences

A

Cell 2 -> Cell 1 Cell 1 -> Cell 2

• 0.04

0.04 200 200

Spikes/bin Time (s)

B

0.02

• 0.02

Cell 4 -> Cell 3 Cell 3 -> Cell 4 200 200

Spikes/bin Time (s)

Signal correlation

Saturday, March 5, 2011

Inhibitory interaction Excitatory interaction

In the inferior temporal cortex, inhibitory interactions exist between neurons with different stimulus preferences

Tamura et al. 2003, 2004

Saturday, March 5, 2011

Recording sites The best stimuli of 55 neurons

! " !# $% $& '" '' &" ##( #%' #%& #') !") !"( !$# !)# !(! !&% %%" "(& $!' $"$ $)* (() #**$ #$)' #(!& !!%& !!$( $!&! ('*& #*!*) ##)"! #"%$$ !*((! !%)$$ !"'$) !(%!# %%"(( ")*'( $&")" !&#!") "#'#%$ "%%(') "($()( $))!"' '""#$! '"(%*% ('#%$" &#)&!' &$#&!) &$("&' ('& * # ! % " $ ) ' ( & #* ## #! #% #" #$ #) #' #( #& !* !# !! !% !" !$ !) !' !( !& %* %#

Cell Most of the functional interactions were observed in adjacent neuron pairs

Cross-correlation

Cell

Normalized center peak height

Vertical connections

(1485 pairs)

Saturday, March 5, 2011

25 24 9 16k 15k 1 19 7k 1 9 15 302 1 17 28 18k 52k 2 4 32 4k 51k 9 24 25 15k 16k 1 19 7k 1 9 15 302 1 17 28 18k 52k 2 4 15 32 4k 51k

200805271733 (28 cells => 378 pairs)

Shank-1 Shank-2 Shank-3 Shank-4 Shank-5 Shank-6 Shank-7 Shank-8 15

Horizontal connections

We found short-range and long-range horizontal connections

Saturday, March 5, 2011

Inhibitory interaction Excitatory interaction

In the inferior temporal cortex, inhibitory interactions exist between neurons with different stimulus preferences

Tamura et al. 2003, 2004

Saturday, March 5, 2011

Representation of visual object information in the inferior temporal (IT) cortex

Saturday, March 5, 2011

How wide is necessary? widely distributed?

• r

locally distributed?

Saturday, March 5, 2011

“The representations are widely distributed”

Haxby et al., Science 293:2425-2430, 2001

Saturday, March 5, 2011

Tsunoda et al., Nat Neurosci 4:832-838, 2001

Complex objects are represented by combinations of feature columns

1 mm

Saturday, March 5, 2011

Adjacent neurons belong to independent networks

Yoshimura et al., Nature 433:868-873, 2005

Saturday, March 5, 2011

65 40 1s 1 64

A Cell 1

Spikes/s 67.5 80

B Cell 2 C Cell 3

1s 1s 40 Spikes/s 60 Spikes/s 1 64 1 64

下側頭葉

0.2

• 0.2 0

1.5

cell 1 cell 2 cell 3

6 ch 1 ch 2 ch 3 ch 4 ch 5 ch mv ms

Tamura et al. 2005

Responses of 3 adjacent neurons to a stimulus set

TE

Saturday, March 5, 2011

Neurosci Res 52:311-322, 2005

Adjacent neurons do not always share stimulus preferences

Saturday, March 5, 2011

How wide is necessary?

We investigated this issue by quantifying single-trial responses with a linear decoding analysis

Saturday, March 5, 2011

I am sorry, I removed 2nd part of my talk

Saturday, March 5, 2011

The representation is locally distributed

Saturday, March 5, 2011

Spike trains of 55 simultaneously recorded neurons

Cell Time (sec) 50

Saturday, March 5, 2011

Collaboration between computational neuroscientists and neurophysiologists (biologists)

http://images.jsc.nasa.gov/lores/S68-49301.jpg

spike trains from multiple neurons

Saturday, March 5, 2011