Attention Reverses the Effect of Prediction in Silencing Sensory Signals

Abstract

Predictive coding models suggest that predicted sensory signals are attenuated (silencing of prediction error). These models, though influential, are challenged by the fact that prediction sometimes seems to enhance rather than reduce sensory signals, as in the case of attentional cueing experiments. One possible explanation is that in these experiments, prediction (i.e., stimulus probability) is confounded with attention (i.e., task relevance), which is known to boost rather than reduce sensory signal. However, recent theoretical work on predictive coding inspires an alternative hypothesis and suggests that attention and prediction operate synergistically to improve the precision of perceptual inference. This model posits that attention leads to heightened weighting of sensory evidence, thereby reversing the sensory silencing by prediction. Here, we factorially manipulated attention and prediction in a functional magnetic resonance imaging study and distinguished between these 2 hypotheses. Our results support a predictive coding model wherein attention reverses the sensory attenuation of predicted signals.

Predictability of what or where reduces brain activity, but a bottleneck occurs when both are predictable

“…In a study that manipulated the frequency of left/right screen-side presentations of simple visual stimuli, presentation of stimuli on the unexpected side resulted in increased activity in V1 in a condition demanding no attention or evaluation of the stimuli (Kok et al, 2012). In another study where auditory tones predicted presence or absence of visual stimuli (den Ouden et al, 2009), surprising events --either unexpected presentations or omissions of visual stimuli --resulted in increased activity.…”

“…As noted previously (Summerfield & Egner, 2009) differentiating the effects of prediction and attention is quite difficult. However, Kok et al (2012) demonstrated it is possible to differentiate the two by showing that predictable stimuli (in visual cortex) tend to produce less activity when the stream is unattended, but greater activity when the stream is attended to. Thus, whether a certain cortical system shows reduced or increased activity to predictable stimuli (vs. unpredictable) may itself depend on attentional state.…”

“…Satisfied -i.e., correct --predictions are associated with reduced prediction errors, and lower activity in sensory regions (Feldman & Friston, 2010;Kok, Rahnev, Jehee, Lau, & de Lange, 2012). Predictions may be instantiated and evaluated via interactions between frontal regions and sensory systems (Bar et al, 2006;Summerfield et al, 2006), or via interactions between sensory regions (den Ouden, Friston, Daw, McIntosh, & Stephan, 2009).…”

“…We therefore expected that being able to predict location transitions would serve as an endogenous cue, resulting in increased activity in regions associated with recruitment and directing of spatial attention, specifically the bilateral intraparietal sulcus and the frontal eye fields (see Corbetta &Shulman, 2002, andalso Szczepanski &Kastner, 2013, for recent review). At the same time, the greater proportion of correct predictions should translate in reduced activity in visual cortex when location is predictable (~ V1; Kok et al, 2012).…”

“…Consistent with this view, Bubic et al (2011) found that different neural systems are involved in evaluating predictions regarding object identity, location, and presentation time. Other studies have shown that odd-ball (unexpected) auditory or visual events produce patterns corresponding to surprise or prediction error in respective sensory cortices (Kok et al, 2012;Mustovic et al, 2003;Todorovic, van Ede, Maris, & de Lange, 2011).…”