Decoding time-resolved neural representations of orientation ensemble perception.

Area of Science:

• Cognitive Neuroscience
• Computational Vision
• Human Perception

Background:

• The visual system rapidly processes summary statistics of visual elements.
• Previous research indicates ensemble perception occurs within 50-200ms, but the precise timing of accurate representation remains unclear.
• Gaps exist in understanding time-resolved neural representations and their behavioral correlation during ensemble perception.

Purpose of the Study:

• To investigate the temporal dynamics of neural representations during orientation ensemble perception.
• To quantify the representational strength of visual ensembles and correlate it with behavioral performance.
• To determine when the visual system forms an accurate ensemble representation linked to perception.

Main Methods:

• Conducted orientation ensemble discrimination tasks with human observers.
• Utilized electroencephalogram (EEG) recordings to capture neural activity.
• Employed an inverted encoding model to decode orientation representations from EEG signals over time.

Main Results:

• An inverted encoding model revealed stronger representations of the average orientation between 400-700ms.
• Neural orientation representations estimated from EEG signals at 600-700ms significantly correlated with individual differences in perceived average orientation.
• Demonstrated a gradual refinement of ensemble orientation representation over several hundred milliseconds.

Conclusions:

• While initial ensemble processing is rapid, accurate ensemble representation develops gradually.
• The visual system refines ensemble orientation perception over hundreds of milliseconds for improved accuracy.
• Neural representations around 600-700ms are critical for accurate individual perception of average orientation.