Abstract
Alpha oscillations are known to impair detection of visual stimuli, but it is unclear if this is due to increased guess rate or decreased fidelity of the mental representation. Here we estimated quality and guess rate as a function of pre-stimulus alpha oscillations using an orientation detection task. In the current study, 21 participants performed a task that consisted of a target pointing in one direction followed by a backward mask. Target orientation was pseudo-randomly chosen from a list of 24 predefined, evenly spaced orientations. Following the mask offset and a 500 ms delay, a response screen appeared where participants used the mouse to rotate the pointing stimuli so that it matches the orientation of the target. Errors were quantified as the difference between the target orientation and the orientation of participants’ response stimuli. During the task, EEG was recorded and the power of alpha was calculated on each trial time-locked to the onset of the target. A median split of alpha power prior to target onset was used to separate the trials into high and low alpha power. The errors of each participant on high and low alpha trials were fit to a standard mixture model to get the parameter values g (guess rate) and σ (precision). We found that g was significantly greater on trials with high alpha power compared to low alpha power. In addition, the σ did not differ significantly between low and high alpha trials. These results indicate that random fluctuations in alpha power can influence the biasing of perception but not the precision of mental representations.