Event-related brain potentials dissociate visual working memory processes under categorial and identical comparison conditions.

Event-related potentials (ERPs) have been successfully employed to examine the functional and neuronal characteristics of working memory processes. In the present study, we examined the ERP waveforms in a delayed matching task to examine the cognitive processes underlying category and identity comparison and the effects of stimulus complexity. Subjects had to decide whether two visual stimuli are (a) physically identical (identical comparison condition, IC) or (b) identical, irrespective of their orientation (categorial comparison condition, CC). The stimuli were structured five-point patterns, which varied in complexity. For the ERPs elicited during the 1500 ms retention interval, the following pattern of results was obtained: Stimuli in the CC-condition elicited larger P300 components than in the IC-condition. In the IC-condition, the P300 was followed by a broadly distributed negative slow wave. Moreover, complex patterns elicited a posteriorily distributed negativity at 350 ms (N350), whereas the less complex patterns gave rise to a fronto-centrally distributed slow wave that started around 500 ms. These results suggest that S1 was more elaborately processed in the CC-condition, while the more complex figures were associated with an early classification process during the retention interval.

Does mental rotation require central mechanisms?

Four reaction time experiments examined the mental rotation process using a psychological refractory period paradigm. On each trial, participants made speeded responses to both a tone (S1) and a rotated letter (S2), presented with varying stimulus onset asynchronies (SOAs). If mental rotation of the stimulus letter can proceed while central mechanisms are busy with S1, then the effect of orientation should decrease substantially with decreasing SOA. Contrary to these predictions, the effect of orientation was nearly constant across SOAs, suggesting that mental rotation cannot effectively proceed without help from central mechanisms. These results support the conclusion that mental rotation requires access to a single-channel mechanism and must therefore be performed serially with other operations requiring the same mechanism.