The stimuli are illustrated as insets at the top: for bars and dot patterns, the white square shows the area (1 cm×1 cm) across which the stimuli are scanned; the gray region illustrates the stimulus extending outside of the stimulation area. The left of each raster is the direction of motion of the stimulus. To the right of each raster is the mean firing rate evoked by stimuli moving in each direction. This neuron produced the most robust response to stimuli moving at approximately 20°, regardless of whether the stimuli were bars, dot patterns, or random dot displays. For random dot displays, the direction tuning increased dramatically with increases in the motion coherence (we show responses at only three levels of coherence for the sake of clarity). The “burstiness” of the response to dot patterns likely reflects individual dots moving across the neuron's hotspot.

(A) Distribution of the direction tuning index, DI, derived from the responses of peripheral afferents and of neurons in areas 3b, 1, and 2 to bars scanned across the fingertip. The responses of individual afferents were not sensitive to the direction of stimulus motion, whereas the responses of many cortical neurons were. The DIs obtained from neurons in area 1 (mean ± s.e.m.: 0.25±0.02) were significantly higher than those obtained from neurons in area 3b (s.e.m.: 0.19±0.02) (t(178) = 2.1, p<0.05). (B) Cumulative distributions of DI obtained from dot patterns scanned across the fingertip. Direction tuning was stronger in area 1 (0.18±0.02) than it was in area 3b (0.13±0.01) (t(138) = 2.1, p<0.05). Furthermore, tuning was weaker for dot patterns than for bars. (C) Cumulative distributions of DI obtained from random dot displays delivered to the fingertip. Direction tuning was stronger in area 1 (0.21±0.03) than it was in area 3b (0.11±0.02) (t(32) = 1.9, p = 0.07). (D) DI as a function of the motion coherence of the random dot displays for neurons in areas 3b (red), 1 (cyan), and 2 (black). The direction tuning of area 1 neurons increased monotonically with motion coherence, whereas that of area 3b neurons did not. Area 2 neurons exhibited tuned responses only when the motion signal was highly coherent (>70%). (E) Preferred direction measured from responses to dot patterns (ϕ_d) versus preferred direction measured from responses to bars (ϕ_b) for neurons that yielded a significant DI for bars, dot patterns, and random dot displays (no neurons in area 3b and only two neurons in area 2 met this criterion). The circular correlation between ϕ_d and ϕ_b was 0.69 overall (p<0.01). 0° denotes left to right motion, 90° proximal to distal, 180° right to left, and 270° distal to proximal. (F). Preferred direction measured from responses to random dot displays at 100% coherence (ϕ_r) versus that measured from responses to dot patterns (ϕ_d) for neurons significantly tuned for all three types of stimuli. The circular correlation between ϕ_d and ϕ_r was 0.94 (p<0.01). Most neurons yielded comparable preferred directions to the two types of stimuli.

(A) Direction selectivity index (DI) as a function of amplitude. (B) Preferred direction obtained from bars of amplitude 700 µm (ϕ_700µm) versus that obtained from bars of amplitude 300 µm (ϕ_300µm). (C) DI as a function of scanning speed. Preferred direction at a non-preferred speed (ϕ_np) versus that at the preferred speed (ϕ_p) for every combination for which direction tuning was significant. Different neurons were more strongly direction-tuned at different speeds with a tendency for neurons to be more strongly tuned at higher speeds (see Figure S4). For the neurons whose direction tuning changed as a function of scanning speed, the increase in tuning could be attributed to an increased firing rate in the preferred direction, to a decreased firing rate in the anti-preferred direction, or to both. A similar modulation of tuning strength by speed is also observed in primary visual cortex.

The neurometric functions were derived from the responses of the neurons that exhibited significant direction tuning to bars, dot patterns, and random dot displays. A positive Δϕ denotes a clockwise rotation. As can be seen, the direction information carried by these neurons can account for the perceived direction of the stimuli. (D) Psychometric and neurometric functions obtained for the left-right discrimination task with random dot displays varying in coherence. Again, the responses of directionally selective area 1 neurons could account for perception.

(A) Orientation selectivity index (OI) versus direction selectivity index (DI) for neurons in areas 3b (crosses), 1 (circles), and 2 (triangles). Red symbols represent neurons that yielded significant OIs, blue symbols represent neurons that yielded significant DIs, and green symbols represent neurons for which both indices were significant; black symbols represent neurons for which neither index was significant. (B) Responses of a neuron that yielded a high OI and a low DI (marked as a B in the upper left quadrant of the scatterplot); red trace is a fitted von Mises function (circular Gaussian); (C) responses of a neuron that yielded a high DI and a low OI (marked as a C in the lower right quadrant); blue trace is a fitted von Mises function; (D) responses of a neuron that yielded intermediate values for both the OI and the DI (marked as a D around the center); green trace is the product of two von Mises functions; (E) Distribution of the angular difference between the preferred direction (measured from dot patterns) and the preferred orientation (measured from indented bars) of neurons in area 1. For orientation, 90° and 270° denote bars parallel to the long axis of the finger, and 0° and 180° denote bars orthogonal to the long axis of the finger. The preponderance of neurons that are sensitive to orientation and direction of motion respond to contours moving in a direction perpendicular to their orientation.