Abstract
V1 neurons respond to simple stimuli like bars and gratings, as well as to complex stimuli such as two-dimensional Hermite patterns (Victor et al., 2006). Here we used two-photon imaging (GCaMP5) to compare responses of V1 superficial layer neurons to simple gratings and complex Hermite patterns in two awake macaques. We recorded at two depths (150μm and 300μm) from 850 × 850 μm2 windows at 4–5° eccentricity. The stimuli were either high-contrast (0.9) drafting (2 cycles/sec) Gabor gratings at 12 orientations, 6 SFs (0.25–8 cpd), and 3 sizes (>= 1 octave, for maximal responses and least surround suppression), or Hermite patterns including vignette, checkboard, and circular/dartboard-like types, at 3 sizes and 6 orientations if possible for maximal responses. V1 neurons were found to respond to either gratings only, Hermite patterns only, or both stimuli. There were more Hermite neurons and less grating neurons at more superficial depth at 150μm than at 300μm (Hermite neurons increased from 18% to 34%; Grating neurons reduced from 64% to 38%). Hermite neuronal responses were at least as strong as grating neuronal responses. More than half (55%) Hermite neurons were activated by circular/dartboard-like patterns, which increased sharply from 300μm to more superficial 150μm depth (from 42% to 63%). Hermite neurons preferred lower SFs than grating neurons (1.25 vs. 2.45 cpd), overlapped with neurons unturned to orientation, and showed some clustering. These results suggest that selectivity to complex stimulus patterns may be an emerging property within V1. The Hermite neurons, mostly clustered in low-SF domains and being orientation unselective, may form a specialized neuron subpopulation for feature integration in V1. The large number of neurons responding to circular/dartboard patterns may carry biological significance for monkey’s survival (e.g., finding round-shaped fruits or faces).