Attentional effort modulated by task difficulty

It has been assumed that stimulus discrimination in a visual task is performed with fixed attentive effort. Here we show that attention to the same pair of stimuli can be modulated by varying the task difficulty when a task requires the discrimination of only a small number of different stimuli. We used a matching-to-sample paradigm, where a test stimulus is presented after a sample stimulus. When both stimuli Gabor gratings have identical orientations ("matching" trial) the required response is different from when they have different orientations ("non-matching" trial). The task difficulty was manipulated by changing the orientation difference between sample and test stimuli for non-matching trials. Difficult non-matching probe trials were embedded within an easy block of trials (easy environment), and vice versa for easy probe trials. Detectability (d') differences for the same pairs of stimuli (probe trials) in the two environments were calculated as a measure for change in attentional effort, regardless of changes in likelihood ratios (beta). Our results show an increase in d' during the difficult task, for both types of probe trials, in paradigms that contained a small number of stimulus combinations. Thus a modulation in attentional effort along a single discrimination dimension is revealed. However it is restricted by the number of stimulus combinations, due to the limited capacity of the attention available for each stimulus combination.