Decoding the activity of neuronal populations in macaque primary visual cortex

Area of Science:

• Neuroscience
• Computational Neuroscience
• Vision Science

Background:

• Visual function relies on accurate neuronal signals from the visual cortex.
• Combining information from multiple neurons enhances accuracy due to individual neuronal response variability.

Purpose of the Study:

• To investigate the reliability of information extracted from simultaneous recordings of neuronal populations in the macaque primary visual cortex.
• To evaluate a decoding framework for inferring visual stimuli from population activity patterns.

Main Methods:

• Developed and tested a decoding framework using linear combination of neuronal responses.
• Compared a simple parametric decoder (assuming neuronal independence) with a sophisticated empirical decoder that accounts for correlated variability.
• Assessed performance for orientation estimation and discrimination tasks.

Main Results:

• The empirical decoder, incorporating correlated neuronal variability, significantly outperformed the parametric decoder.
• The structure of neuronal response distributions contains crucial information for effective sensory decoding.
• Population activity patterns provide a reliable basis for inferring visual stimuli.

Conclusions:

• Neuronal response structure, including correlated variability, is vital for accurate sensory decoding.
• Sophisticated decoding models that leverage this structure can improve the inference of visual stimuli from population activity.
• Understanding population coding enhances our knowledge of how perceptual decisions are informed by neural signals.