2005 Fiscal Year Final Research Report Summary
Neural mechanisms in early visual stages representing angles and junctions embedded within contour stimuli.
| Project/Area Number |
14380371
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (B)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
Neuroscience in general
|
|---|
| Research Institution | National Institute for Physiological Sciences (2004-2005)
Okazaki National Research Institutes (2002-2003) |
|---|
Principal Investigator |
ITO Minami National Institute for Physiological Sciences, Department of Information Physiology, Associate Professor, 生体情報研究系, 助教授 (20311194)
|
|---|
| Co-Investigator(Kenkyū-buntansha) |
KOMATSU Hidehiko National Institute for Physiological Sciences, Department of Information Physiology, Professor, 生体情報研究系, 教授 (00153669)
OGAWA Tadashi National Institute for Physiological Sciences, Department of Information Physiology, Assistant Professor, 生体情報研究系, 助手 (50311197)
|
|---|
| Project Period (FY) |
2002 – 2005
|
| Keywords | Visual Cortex / Contour line / Angle / Primate / Area V2 / Area V1 |
|---|
| Research Abstract |
This study aimed at revealing the neuronal mechanisms representing angles embedded within contour stimuli in the early visual stages. First, we recorded from the superficial layer of area V2, while monkeys performed the fixation task. Stimuli were made by combining two line components, large enough to pass through the classical receptive fields. We found that (1)fairly large number of neurons showed selective responses to these angle stimuli in V2. Angle width of the optimal angles distributed from 30 to 180 degree. (2)Both or either half-line components were explicitly represented in neurons' direction profile to individual half-lines. Second, we asked whether a relatively simple linear-summation model explains the observed angle selectivity. We assumed two excitatory subunits representing two optimal line components and one inhibitory subunit. Outputs of these subunits were summed in linear manner, then were multiplied by a gain, and finally went through a non-linear filter of rectification. In majority (73/91), optimized model explains all responses to 66 angles and 12 half-lines well (p<0.01 Bootstrap test). These results suggest that the angle selectivity is based on an appropriate conversion of excitatory inputs, which represent optimal half-line components, and inhibitory inputs, which sharpen the angle selectivity due to suppression to non-optimal line components included in angles. Such inhibition and rectification were enough to explain fine tuning to the angle selectivity. Third, we asked the role of V1, which is the major source of visual inputs to area V2. By the similar stimulus set, we found that the optimal angle width of 60 and 120 degree was rarely observed in V1. Some V1 neurons responded to only short line segments. Therefore, V1 neurons may contribute to angle representation in a restricted manner, while V2 neurons represent various angles.
|
