Neuronal representation of subjective shapes in area V4. Cox, M.A. , Schmid, M.C. , Peters, A.J. , Saunders, R.C. , Leopold, D.A. & Maier, A. Section on Cognitive ...
Neuronal representation of subjective shapes in area V4 Cox, M.A.1,2, Schmid, M.C.1,3, Peters, A.J.1,4, Saunders, R.C.1, Leopold, D.A.1 & Maier, A.1,2 1Section on Cognitive Neurophysiology and Imaging, Laboratory of Neuropsychology, National Institute of Mental Health, NIH, Bethesda, MD 20852, USA 2Department of Psychological Sciences, Vanderbilt University, Nashville, TN 37212, USA 3Ernst Strüngmann Institute, Max Planck Society, Frankfurt, Germany 4Neuroscience, University of California-San Diego, La Jolla, CA 92093, USA
Motivation Subjective contour illusions demonstrate the visual system’s inductive mechanisms for contour and surface perception and thus are ideal stimuli for studying the neural underpinnings of these perceptual processes.
CICN
Vanderbilt University
Single-Unit Activity
Receptive Field Analysis
Spiking activity for subjective shapes is elevated compared to control.
Receptive field position predicts subjective shape sensitivity in V4.
Kanizsa’s Triangle
Single-Unit Raster Plots
Mapping of Receptive Field Center via Reverse Correlation:
Monkey B
Monkey F
Visual Space: lower right visal quadrant 0
0
2
4
-2
Neural responses to subjective contours were first measured in macaque V2 by von der Heydt et al. in 1984 [1]. V2
50
0
0
500
Time (ms)
1000
2) Presentation of simple stimulus at various locations in visual space
Modulation Across Population
100
50
V4
0
0
Time (ms)
500
Neurons in both monkeys show more spiking activity evoked by the subjective shape stimulus than by the rotated control. The subjective shape response often showed an oscillatory temporal modulation as in the left example here.
Evidence suggests that subjective contour responses in early visual cortex are a result of feedback from extrastriate areas, such as V4.
Population Summary Representative Multi-Unit Responses Across Array
Modulation Across Population
P V
Modeling: V1 and V2 receptive fields are too small and conduction delays too long for local extraction of subjective contours [3].
V4?
Human fMRI: greater perceptual modulation for subjective stimuli found in higher-tier visual areas than lower-tier areas [4].
P
Broadband Signal 01.Hz-24kHz
Multi-Unit Activity 300Hz-12kHz Rectified
Microelectrode Array electrode length = 1.5mm electrode pitch = 400μm electrode tip radius = 3-5μm
Short Break & Juice Reward
0
(3,-3)
20
10
t-score window 0
500
1000
Spectral Analysis: Between Conditions 200
Frequency
200
Control #1
Control #2
1000
Time (ms)
1000
−100
References
20
100
+
0 0
0
−200
100
0
100
Fractional Power Difference =
-0.5 0
500
Time (ms)
1000
These findings suggest that neurons in area V4 play an important role in subjective contour perception. We plan to further investigate V4 responses to subjective shapes using combined optical imaging and laminar electrophysiology.
Visually Evoked Potential
0.5
100
Time (ms)
V4 neurons with receptive fields covering the center of the stimulus had elevated spiking activity when the inducers created a subjective shape.
Time (ms)
Broadband LFP power is stronger for subjective shape than control.
0
(3,-3)
(3,-3)
0.08
Summary and Future Directions
Local Field Potential
0
0.12
0.04
200
Illusion #2
(0,0) (0,0)
30
0
1000
Time (ms)
Spectral Analysis
Control Stimuli
Stimuli always presented at same location in visual space: bottom right inducer at (3,-3) visual degrees.
(0,0)
Monkey B: Grand Average of All Channels and Sessions, n=767
Illusion #1
0.16
Location of Receptive Field Centers in Visual Space
0
Voltage (μV)
Stimulus Presentation (1s)
Subjective Stimuli
10
0
Location of Receptive Field Centers in Visual Space
40
Difference (μV)
Fixation (1s)
Local Field Potential 1-300Hz
20
Frequency
Experimental Paradigm
Location of Receptive Field Centers in Visual Space
Units with a significant and positive t-score.
Single-Unit, n=1698 Both the multi-unit and single-unit averages showed significant, sustained elevation of neuronal activity for the subjective stimulus compared to the control. Additionally, subjective shapes elicited a prominent slow oscillation consisting of 2-3 cycles of peak spiking at ~200 ms intervals.
Frequency
Fixation Spot
Single-Unit Activity via Spike Sorting
Units with a non-significant t-score.
Units with a significant and negative t-score.
Mean Spike Rate (impulses/sec)
D
Percent Change from Baseline
Methods V
43%
Multi-Unit 30
Visual Presentation
36%
Grand Averages Across Both Monkeys
Do V4 neurons integrate subjective stimuli?
0 10 T-score, Shaded Area Significant p
