The current study examined the interface of visual perception and attention in schizophrenia using two RSVP tasks: a single target task to assess group differences in visual perception and a dual-target task to elicit the AB effect. The single target task showed the expected perceptual deficits in perceiving a brief target within a stream of distractors in patients. The conditional probability data from the dual target task revealed significant performance deficits in the patients but interpretation of this difference is ambiguous because performance on the task involves both perceptual and higher order, attentionally-modulated processes. Using data from both the single and dual target tasks, we examined the suppression ratio to assess the AB effect after accounting for group differences in perception. This ratio revealed group differences in performances both within and outside the typical AB window. These findings suggest that visual processing deficits in schizophrenia patients exist at the interface between perception and higher order processing.
Both patients and controls showed a decrease in performance after lag 1 and both showed an improvement after lag 4, indicating the presence of the AB. However, patients’ performance continued to diminish from lags 2 – 4, whereas the controls’ performance remained stable across these lags. This continued decrease in performance coupled with the significant impairment in performance relative to controls across lags 2-8 indicates that the patients are demonstrating a deeper AB effect. Additionally, the groups’ performance also differed before 200 ms and after 500 ms, with controls showing no performance degradation at either lag 1 or lag 12 whereas patients showed a significant impairment relative to controls at both of these lags. This pattern of findings suggests that attentional processes are affecting the patients’ performance even outside the AB. The role of attention in visual processing has been widely studied, and attentional processes have been implicated in both the maintenance of a percept in visual short-term memory and in the transfer of a percept from iconic memory to visual short-term memory (for review: Cowan 2005
). Given that iconic memory has been shown to be unimpaired in schizophrenia (Green et al., 2010
; Hahn et al., 2010
), the observed impairment at lags 1 and 12 in the schizophrenia patients is possibly attributable to impaired transfer of the percept from iconic memory due to the additional attentional demands of the dual-target task relative to the single-target task.
As noted above, interpretation of AB in healthy individuals is subject to a multitude of theories that attempt to explain the phenomenon. These theories include a two-stage model emphasizing a processing bottleneck that impairs consolidation of visual stimuli in VSTM (Chun & Potter, 1995
) and newer, distractor-driven models that emphasize selective attentional allocation based on the nature of the stimuli and where they fall in the stream of stimuli (Di Lollo et al., 2005
; Olivers & Meeter, 2008
). Although the various models differ in specifics, they all propose that the AB paradigm disrupts processing of visual stimuli at a level higher than basic early visual perception and involves the transition of visual processing from basic perceptual mechanisms to attentionally-modulated higher order cognitive mechanisms.
Previous studies yielded mixed findings regarding AB deficits in schizophrenia, and the current study was an attempt to overcome limitations of these previous studies through the use of large samples and a control procedure for visual perception. The results of the current study provide robust support for the exaggerated AB effect in schizophrenia suggested by the previous studies. The previous studies that found an exaggerated AB effect in schizophrenia were not designed to separate higher order processes from early visual perception deficits (Li et al., 2002
; Wynn et al., 2006
). In contrast, the methods employed in the current study allowed us to investigate this distinction directly. Our findings indicate that the AB effect is exaggerated in schizophrenia even after accounting for group differences in perception. Furthermore, the only other published study that accounted for perceptual deficits did not employ a sample large enough to detect specific impairments across different lags during the AB, and it yielded mixed results (Cheung et al., 2002
). The current findings resolve this ambiguity by providing evidence for impaired performance in schizophrenia across lags throughout, and beyond, the typical AB timeframe. Overall, the results of the current study indicate a problem in schizophrenia at the perception/attention interface on an index that is designed to account for deficits in perception. This problem might be attributable to deficits in control of attention (the basic ability to guide attention in response to internal representations), which is known to be dysfunctional in schizophrenia (Nuechterlein et al., 2009
The findings from the suppression ratio helped to resolve three possible patterns of group differences in the AB effect. First, no group differences in SR would have indicated that the more pronounced AB in patients was due solely to deficits in visual perception that are controlled by the index. Second, group differences in SR only during the AB time window would have indicated that the more pronounced AB in patients was due solely to deficits in the interface between perception and attention. Finally, group differences in SR across all lags (within and outside the AB window) would have indicated that the more pronounced AB in patients are due to deficits in the interface between perception and attention as well as general attentional differences. Our results match the third possibility most closely, making it difficult to disentangle effects from the perception/attention interface versus a general attentional deficit that acts on perception.
Resolving this ambiguity between differences at the interface and the attentional demands of the dual-target task could be accomplished in future studies through refinement of the AB procedure (e.g., adjusting the task demands in order to equalize performance across the groups at lags outside of the traditional AB timeframe), and through the collection of concurrent electroencephalographic (EEG) data. EEG methodology, specifically the measurement of event-related potentials, provides the temporal resolution and theoretical underpinning necessary to differentiate between performance deficits associated with low-level, bottom up processes, higher-level, intrinsic processes, and the interface between the two. Building upon the already considerable work in schizophrenia using event-related potentials (e.g., P1, N1, P300) subsequent studies of the perception/attention interface using the AB paradigm and concurrent EEG would likely be able to further isolate deficits specific to different stages of visual processing.