Abstract
The ability to consciously recognise visual objects is crucial for adaptive behaviour and survival. Conscious access to visual objects has been studied using the Attentional Blink (AB), in which two targets (T1 and T2) are embedded with visual masks in a rapid serial visual presentation (RSVP). In the AB, the ability to detect T2 is reduced when presented 200-500ms post T1. Research using functional Magnetic Resonance Imaging (fMRI) has proven useful to identify the underlying brain mechanisms of conscious access. Given the challenges inherent to the limited temporal resolution of fMRI, researchers have designed AB-studies in which T1 and T2 targets are selected from image categories known to engage different regions in the visual stream. However, to integrate these findings into a consistent model of conscious access, the variability in detection thresholds across categories needs to be assessed. Here, we presented participants with 48 pictures of objects from several categories in an AB-task. Each picture was presented as T1, and at two different T2-lags (200ms and 700ms post T1). To compare the performance at recalling target objects across categories, we used a factorial ANOVA with T2-lag and object category as factors. We observed main effects of T2-lag (F(1,20)=51.47, p < 0.001) and category (F(7,140)=51.6, p < 0.001), along with an interaction between category and T2-lag (F(7, 140)=27.4, p < 0.001). Beyond the expected AB effect, this means that different object categories exhibit different detection thresholds. We further pooled the categories according to animate and inanimate objects, which are known to vary in their processing speeds. Here, a pairwise t-test revealed a markedly smaller AB-magnitude for animate objects (t=4.51, df=37.29, p < 0.001). These findings indicate a behavioural advantage for animate objects in their representational readouts, advocating for careful consideration of stimulus materials in conscious access research.
Meeting abstract presented at VSS 2017