To see, not to see or to see poorly: Perceptual quality and guess rate as a function of electroencephalography (EEG) brain activity in an orientation perception task

• Published in: The European journal of neuroscience Vol. 55; no. 11-12; pp. 3154 - 3177
• Main Authors: Sheldon, Sarah S., Mathewson, Kyle E.
• Format: Journal Article
• Language: English
• Published: Chichester Wiley Subscription Services, Inc 01-06-2022
• Subjects: EEG; Electroencephalography; Event-related potentials; Orientation behavior; orientation perception; standard mixture model; time‐frequency analysis; variable precision model; Visual perception; Visual stimuli
• DOI: 10.1111/ejn.15445
• EISSN: 1460-9568
• ISSN: 0953-816X

Detection of visual stimuli fluctuates over time, and these fluctuations have been shown to correlate with time domain evoked activity and frequency‐domain periodic activity. However, it is unclear if these fluctuations are related to a change in guess rate, perceptual quality or both. Here we determined whether the quality of perception randomly varies across trials or is fixed so that the variability is the same. Then we estimated how perceptual quality and guess rate on an orientation perception task relate to electroencephalography (EEG) activity. Response errors were fitted to variable precision models and the standard mixture model to determine whether perceptual quality is from a varying or fixed distribution. Overall, the best fit was the standard mixture model that assumes that response variability can be defined by a fixed distribution. The power and phase of 2–7 Hz post‐target activities were found to vary along with task performance in that more accurate trials had greater power, and the preferred phase differed significantly between accurate and guess trials. Guess rate and σ were significantly lower on trials with high 2‐ to 3‐Hz power than low, and the difference started around 250‐ms post‐target. These effects coincide with changes in the P3 event‐related potential (ERP): There was a more positive deflection in the accurate trials versus guesses. These results suggest that the spread of errors (perceptual quality) can be characterised by a fixed range of values. Where the errors fall within that range is modulated by the post‐target power in the lower‐frequency bands and their analogous ERPs.
Performance variability on a visual orientation perception task was best defined by a single distribution. Post‐target 2‐ to 7‐Hz electroencephalography (EEG) activity and the P3 event‐related potential (ERP) varied with task performance in that more accurate trials had greater power and a more positive P3. These results suggest that perceptual quality can be characterised by a fixed range, and where it falls within that range is modulated by the post‐target power in the lower‐frequency bands and their analogous ERPs.