Related Experiment Videos
Do Gabor functions provide appropriate descriptions of visual cortical receptive fields?
1Department of Psychology, Stanford University, California 94305.
Summary
Human spatial vision models suggest Gabor functions, but these complex-valued functions don't fit single-cell data. Alternative receptive fields may better explain visual processing, challenging current theories.
Area of Science:
- Neuroscience
- Computational Vision
- Fourier Analysis
Background:
- Theoretical models of human spatial vision often utilize receptive fields that minimize uncertainty products.
- Gabor functions are proposed as optimal receptive fields based on Fourier analysis principles.
Purpose of the Study:
- To evaluate the suitability of Gabor functions for modeling cortical receptive fields in human spatial vision.
- To explore alternative receptive field functions beyond Gabor functions.
Main Methods:
- Analysis of theoretical models of spatial vision.
- Review of neurophysiological measurements.
- Psychophysical masking data analysis.
Main Results:
- Complex-valued Gabor functions, while minimizing uncertainty, are not directly applicable to single-cell receptive field measurements.
- Alternative metrics for positional and spatial-frequency uncertainty yield biologically plausible receptive field functions.
- Receptive field functions other than Gabor functions provide a better fit to existing neurophysiological and psychophysical data in many cases.
Conclusions:
- Current theoretical and experimental evidence is insufficient to exclusively favor Gabor functions for modeling human visual receptive fields.
- A broader class of receptive field functions should be considered for accurately describing visual processing.