Cortical correlates of human motion perception biases

Area of Science:

• Neuroscience
• Visual Perception
• Computational Neuroscience

Background:

• Human sensory perception is not a veridical representation of the environment.
• Perceptual biases, such as objects appearing slower at low contrast, have been observed but their neural basis remains unknown.
• This suggests a potential disconnect between motion perception and cortical responses.

Purpose of the Study:

• To investigate the neural encoding of speed perception biases in the human cortex.
• To determine where and how these perceptual biases are represented at the neuronal population level.
• To resolve the disconnect between observed motion perception biases and cortical activity.

Main Methods:

• Functional magnetic resonance imaging (fMRI) was employed to study neural populations.
• Decoding, voxel population, and forward-encoding analyses were utilized.
• The encoding of speed and temporal frequency at varying contrast levels was examined.

Main Results:

• Perceptual biases toward slower speeds and higher temporal frequencies at low contrast were identified.
• These biases were found to be encoded in the earliest visual cortical regions.
• The neural encoding in early visual areas accurately matched human perceptual experience.

Conclusions:

• The study resolves the disconnect between cortical responses and human motion perception.
• Perceptual priors, such as a bias for slow speeds, are encoded in early visual cortical areas.
• These findings support the role of early visual cortex in both sensory evidence processing and the encoding of perceptual priors, aligning with Bayesian inference models.