Internal noise in contrast discrimination propagates forwards from early visual cortex

Area of Science:

• Neuroscience
• Visual Perception
• Cognitive Science

Background:

• Human contrast discrimination is limited by neural noise and transduction nonlinearities.
• The precise timing and location of internal neural noise remain debated, with psychophysical models suggesting late processing and neuroimaging studies indicating early processing.

Purpose of the Study:

• To investigate whether perceptually-relevant internal noise originates in early visual areas or later decision-making regions.
• To determine the temporal dynamics of neural noise during contrast discrimination.

Main Methods:

• Recorded electroencephalography (EEG) and magnetoencephalography (MEG) during a two-interval-forced-choice contrast discrimination task.
• Employed multivariate pattern analysis (MVPA) to decode perceptual decisions from evoked brain responses.
• Utilized MEG source space analysis to examine information spread across anatomically defined brain regions.

Main Results:

• Perceptual decisions were successfully decoded from both EEG and MEG signals, even with identical stimuli.
• Above-chance classification of decisions began less than 100 milliseconds (ms) after stimulus onset, indicating early neural noise.
• Classification accuracy increased over time, peaking beyond 500 ms, with initial information in occipital regions spreading to parietal and frontal areas.

Conclusions:

• Neural noise significantly affects sensory signals early in the visual pathway, influencing perceptual decisions.
• Noise impacts sensory processing across multiple stages, from early visual areas to later decision-making regions.
• Early sensory noise may be resolved through mechanisms like Birdsall's linearisation, balancing dynamic range preservation with contrast-invariance.