Figure 1.
Schematic illustration of the research focus and the logic of the experiments undertaken.
(A) A schematic diagram that illustrates motion direction, speed, and axis in the case of an apparent motion stimulus, in which each bulb is lighted sequentially in rightward direction at different times without any actual position changes (no real motion involved but a series of retinotopic activations in space and time, i.e. motion trajectory). The sequential retinotopic activations by flashes can lead to the perception of direction, axis, and speed. (B) The experimental approach. Direction-selective population and single-cell responses to random-dot motion from cat early visual areas 17 and 18 were acquired by the combination of intrinsic-signal optical imaging and single-unit recordings. Arrows inside the figure demonstrate the definition of motion direction that ranges from 0° to 360°. Note that motion axis is the same for dots moving in opposite directions and ranges from 0° to 180°. Examples of results from areas 17 and 18 are demonstrated.
Figure 2.
Stimuli used in this study and the discrimination of areas 17 and 18.
(A–B) 0° oriented sine-wave grating with spatial frequency of 0.5 cycles per degree (cpd) (A) and random-dot field moving towards 0° direction (B). The Fourier power spectra and the power distributions along the spatial frequency dimension of the static grating and dot stimuli are shown below each stimulus diagram. Note that the power spectrum of the static random-dot field exhibits a uniform distribution in the orientation dimension. Most of the stimulus power resides at the spatial frequency of 0.5 cpd for the sine-wave grating and below 3 cpd for the random-dot field. (C) The discrimination of areas 17 and 18. An image of the cortical surface taken under 550 nm green-light illumination is displayed. The area boxed by the red rectangle was the region of interest (ROI) for optical imaging. Polar maps of orientation preference derived from sine-wave grating stimuli with spatial frequencies of 0.75 cpd and 0.15 cpd were constructed respectively, and the white broken line delimits the border between area 17 and 18 accordingly. A, Anterior; L, lateral. Scale bar: 1 mm.
Table 1.
Parameters used for the model simulation.
Figure 3.
Differential population responses to different motion axes.
(A) Differential orientation maps generated by drifting sine-wave gratings with SFs of 0.6 cpd and 0.2 cpd for driving areas 17 and 18, respectively. Blood vessels and marginal regions were masked gray on the maps. Responses of selected regions from areas 17 and 18 (green and orange outlines) were analyzed in (B) and (C). A, Anterior; L, lateral. (B) Differential orientation maps (0°–90°) and the corresponding results of response profile analysis acquired using sine-wave grating stimuli. (C) Differential motion-axis maps (90°–0°) and the corresponding results of response profile analysis obtained by moving random-dot stimuli with different speeds. Colored iso-orientation contours were derived from the orientation preference map for sine-wave grating stimuli and were superimposed on the gray images. The red curves in the plots represent the best fitting cosine functions. The values of the ordinates correspond to the recorded responses () × 104. Note that the scales of y axes are different between low and high speeds. Scale bar: 1 mm.
Figure 4.
The responses of two orientation-selective cells in areas 17 and 18 to different moving directions and speeds of random dots.
(A) Typical polar plots of direction tunings of orientation-selective cells generated using sine-wave grating stimuli. PO, preferred orientation; OSI, orientation-selective index. N = 5 trials for both cells. (B) Polar plots of direction tunings acquired by using random-dot stimuli moving at different speeds. PA, preferred motion axis; ASI, motion-axis selective index. N = 10 trials for both cells. The spontaneous firing rates of the cells are 13 spikes/s and 3 spikes/s in areas 17 and 18, respectively. Error bars represent SEM.
Figure 5.
The simulated responses of orientation-selective cells to moving random-dot stimuli with different speeds.
(A) Polar plots of direction tunings of simulated neuronal responses to drifting sine-wave gratings. PO, preferred orientation; OSI, orientation-selective index. (B) Polar plots of direction tunings of simulated neuronal responses to random-dot stimuli moving at different speeds. PA, preferred motion axis; ASI, motion-axis selective index.
Figure 6.
Differential population responses to different directions of motion.
(A) Image of the cortical vasculature and schematic diagrams of the stimuli used for driving direction-selective responses. The irregular green and orange polygons define the selected regions of areas 17 and 18 for further analysis, respectively. Red arrows in the stimulus diagrams indicate the moving directions. A, Anterior; L, lateral. Scale bar: 1 mm. (B) Differential direction maps (0°–180°) and the corresponding results of response profile analysis acquired using stimuli of drifting sine-wave gratings. (C) Differential direction maps (0°–180°) and the corresponding results of response profile analysis obtained by moving random-dot stimuli with different speeds. Colored iso-direction contours were derived from the direction preference map for sine-wave grating stimuli and were superimposed on the gray images. The red curves in the plots represent the best fitting cosine functions. The values of the ordinates correspond to the recorded responses () × 104.
Figure 7.
Direction preference maps calculated by vector summation were not changed with moving speed of random-dot stimuli.
(A) Image of the cortical vasculature. The green and orange polygons indicate regions of areas 17 and 18 selected for analysis, respectively. A, Anterior; L, lateral. Scale bar: 1 mm. (B) Direction preference maps calculated by vector summation acquired using sine-wave grating stimuli in areas 17 and 18. Hues in the color maps indicate direction preferences. (C) Differential direction maps activated by four different pairs of directions of moving sine-wave gratings and random dots with different speeds. (D–E) Direction preference maps calculated by vector summation acquired using random-dot stimuli with different speeds in areas 17 (D) and 18 (E). Angular differences between direction preference maps of sine-wave grating and of random-dot stimuli with different speeds are shown in the histograms. Computed from the angular-difference histograms of 5°/s, 30°/s, and 60°/s,the percentages of pixels with angular difference between −60° and 60° amount to 73%, 78%, and 75% in area 17 (D) and 78%, 82%, and 80% in area 18 (E), respectively.
Figure 8.
Direction preference maps calculated by vector maximum were changed with moving speed of random-dot stimuli.
(A) Image of the cortical vasculature. The green and orange polygons indicate regions of areas 17 and 18 selected for analysis, respectively. Note that the selected regions are the same regions analyzed in Figure 7. A, Anterior; L, lateral. Scale bar: 1 mm. (B) Direction preference maps calculated by vector maximum acquired using sine-wave grating stimuli in areas 17 and 18. Hues in the color maps indicate direction preferences. (C–D) Direction preference maps calculated by vector maximum acquired using random-dot stimuli with different speeds in areas 17 (C) and 18 (D). Angular differences between direction preference maps of sine-wave grating and of random-dot stimuli with different speeds are shown in the histograms. Computed from the angular-difference histograms of 5°/s, 30°/s, and 60°/s, the percentages of pixels with angular difference between −45° and 45° amount to 71%, 49%, and 40% in area 17 (C) and 69%, 64%, and 65% in area 18 (D), respectively.
Figure 9.
Direction differential maps and response profile analyses from a posterior part of area 17.
(A) Image of the cortical vasculature. The white broken line delimits the border between area 17 and 18. The green polygon indicates the posterior part of area 17 selected for analysis. A, Anterior; L, lateral. Scale bar: 1 mm. (B–C) Differential direction maps (0°–180°, B; 90°–270°, C) and the corresponding results of response profile analysis obtained by moving random-dot stimuli with different speeds. Colored iso-direction contours were derived from the direction preference map for sine-wave gratings and were superimposed on the gray images. The red curves in the plots represent the best fitting cosine functions. The values of the ordinates correspond to the recorded responses () × 104.
Figure 10.
Summary of the population responses to random-dot motion with different speeds in areas 17 and 18.
(A) Angular differences between the fitted orientations of differential motion-axis maps and the motion axes of the moving-dot stimuli were plotted against speed. The red dotted curves represent the quadratic polynomial fits. (B) Angular differences between the fitted directions of differential direction maps and the moving directions of the random-dot stimuli. Results were calculated through all speeds tested and averaged across 5 cats studied using intrinsic optical imaging. Error bars represent SEM.
Figure 11.
Responses of two direction-selective cells in areas 17 and 18 to different moving directions and speeds of random dots.
(A) Typical polar plots of direction tunings of direction-selective cells generated using sine-wave grating stimuli. PD, preferred direction; DSI, direction selective index; PO, preferred orientation; OSI, orientation selective index. N = 6 and 5 trials for the cells in areas 17 and 18 respectively. (B) Polar plots of direction tunings acquired by using random-dot stimuli moving at different speeds. PA, preferred motion axis; ASI, motion-axis selective index. N = 10 trials for both cells. The spontaneous firing rates of the cells are 0 spikes/s and 2 spikes/s in areas 17 and 18 respectively. Error bars represent SEM.
Figure 12.
Summary of the response properties of direction-selective neurons to random-dot motion at different speeds.
(A) Angular differences between the preferred orientations for sine-wave gratings and the preferred motion axes for moving random dots. (B) The influence of speed on motion-axis selective index (ASI). (C) Angular differences between the preferred directions for moving sine-wave gratings and for moving random dots. (D) The influence of speed on direction-selective index (DSI). Error bars represent SEM.
Figure 13.
Response properties of direction-selective neurons to moving dots with size of 0.4° and 0.8° were similar.
(A) Angular differences between the preferred orientations for sine-wave gratings and the preferred motion axes for moving random dots. (B) The influence of dot size on motion-axis selective index (ASI). (C) Angular differences between the preferred directions for moving sine-wave gratings and for moving random dots. (D) The influence of dot size on direction-selective index (DSI). Error bars represent SEM.
Figure 14.
Response properties of simple and complex cells were similar to random-dot motion with different speeds.
(A) Angular differences between the preferred motion axes for moving random-dot field and the preferred orientations for sine-wave gratings. (B) Angular differences between the preferred directions for moving random-dot field and for moving sine-wave gratings. Results were calculated across all speeds tested for simple and complex cells in areas 17 and 18 respectively. Error bars represent SEM.
Figure 15.
Model simulations for direction-selective cells at both single-cell and population levels.
(A–B) Polar plots of direction tunings of direction-selective model cells in areas 17 (A) and 18 (B) to drifting sine-wave gratings and random dots. PD, preferred direction; DSI, direction-selective index; PO, preferred orientation; OSI, orientation-selective index. PA, preferred motion axis; ASI, motion-axis selective index. (C) Simulated motion-axis selective indexes of direction-selective neuronal populations. (D) Angular differences between the fitted orientations of simulated differential population responses to motion axis and the motion axes of the moving-dot stimuli used. The gray broken lines correspond to 0° and 90° angular differences. (E) The influence of dot size on the simulated transition speed. When the dot size increased from 0.2 degree to 6.4 degree, the transition speeds increased from 8.5°/s to 11.5°/s in area 17, and 40°/s to 59°/s in area 18. (F) Simulated direction-selective indexes of direction-selective neuronal populations. (G) Angular differences between the fitted directions of simulated differential population responses to motion direction and the directions of the random-dot stimuli used. The gray broken line corresponds to 0° angular difference. Error bars represent SEM. All population results were averaged from 32 simulated trials, but some error bars were too small to be shown.
Figure 16.
A neuronal mechanism for the processing of random-dot motion in areas 17 and 18.
(A) The illustration of how single-neuron behavior accounts for differential population responses. Diagrams of the stimuli (0°–180°) used for generating differential direction responses are shown at the top. The patterns of the differential response images are similar across different speeds, but not the differential response strengths. The solid circles indicate regions that prefer rightward motion (0°), while areas in dotted circles prefer leftward motion (180°). The length of the arrows in the solid/dotted circles represents the response strength of the underlying direction-selective neurons to different directions. The corresponding direction tuning curves under different speeds for neurons preferring a 0° direction are shown as an example at the bottom. (B) Summary of the behavior of orientation- and direction-selective neurons with orthogonal preferences in the processing of the rightward motion at different speeds. The colored circles represent for orientation- and direction-selective neurons with vigorous responses, while those gray circles indicate less active or silent neurons in the processing of the rightward motion below and above the transition speeds. Markers inside these circles indicate stimulus features encoded by the neurons: bar for orientation, dotted line for motion axis, and arrow for direction. Note that direction-selective neurons in area 18 maintain some responses to rightward direction at high speed.