Evaluating the neurophysiological evidence for predictive processing as a model of perception

Abstract

For many years, the dominant theoretical framework guiding research into the neural origins of perceptual experience has been provided by hierarchical feedforward models, in which sensory inputs are passed through a series of increasingly complex feature detectors. However, the long‐standing orthodoxy of these accounts has recently been challenged by a radically different set of theories that contend that perception arises from a purely inferential process supported by two distinct classes of neurons: those that transmit predictions about sensory states and those that signal sensory information that deviates from those predictions. Although these predictive processing (PP) models have become increasingly influential in cognitive neuroscience, they are also criticized for lacking the empirical support to justify their status. This limited evidence base partly reflects the considerable methodological challenges that are presented when trying to test the unique predictions of these models. However, a confluence of technological and theoretical advances has prompted a recent surge in human and nonhuman neurophysiological research seeking to fill this empirical gap. Here, we will review this new research and evaluate the degree to which its findings support the key claims of PP.

Individuals with autism have no detectable deficit in neural markers of prediction error when presented with auditory rhythms of varied temporal complexity

“…Much recent attention has been turned toward a deficit in predictive coding as a primary neurophysiologic mechanism underlying many of the sensory and other symptoms of ASD (Van de Cruys et al, 2014). Predictive coding is the process by which the brain uses previous experience to predict and influence the processing of incoming sensory input via feedback from top down signals to lower level sensory regions (Walsh, McGovern, Clark, & O'Connell, 2020). Many have proposed a deficit in predictive coding in individuals with ASD, whereby there is diminished contribution of top down predictive mechanisms to lower level sensory areas in the processing of incoming auditory and visual stimuli (Pellicano & Burr, 2012;Van de Cruys et al, 2014).…”

Deep Predictive Learning in Vision

“…For example, although Issa, Cadieu, and DiCarlo (2018) observed an error-signal-like increase in activation for atypical faces in some pIT neurons, these neurons overall had a positive stimulus encoding, with only a relatively small, later, error-like modulation. Furthermore, as discussed below, anticipatory predictions typically closely resemble the subsequent stimulus-driven activity, suggesting a positive, not inhibitory, effect (Cavanagh, Hunt, Afraz, & Rolfs, 2010;Duhamel, Colby, & Goldberg, 1992;Lee & Mumford, 2003;Walsh et al, 2020). However, there are various different ways of reformulating the neural implementation of EE that can avoid some of these issues (Bastos et al, 2012;Spratling, 2008), but perhaps this flexibility renders the framework difficult to falsify (Kogo & Trengove, prediction (minus) actual (plus) 5IB The Bayesian-style explicit error (EE) coding model (Rao & Ballard, 1999;Friston, 2010, Bastos et al, 2012, in a situation where the prediction is clearly erroneous (ball predicted to emerge on right, actually emerges on left).…”

“…To encode both signs of the error (omissions, false alarms) with positive-only spike rates, two separate populations of EE neurons would be required, or a more complicated deviation from tonic firing level scheme. Unambiguous evidence of such EE coding neurons has not been found (Walsh et al, 2020). In contrast, error signals in our proposed framework remain as a temporal difference between the two states of prediction vs. outcome, which enables all connectivity between cortical areas to be excitatory and always represent a positive encoding of either the prediction or outcome.…”

“…2015). In any case, an extensive treatment of the issues with EE is beyond the scope of this paper and has already been aptly covered by Walsh et al (2020) -our goal here is to highlight some of the core differences as a way to clarify the framework by way of contrast, and in relation to available data. First, there are many examples of anticipatory predictive neural firing in the brain.…”

“…This phenomenon is widely cited as evidence in favor of the EE predictive coding framework, consistent with an inhibitory effect of the expectation. Nevertheless, despite various conflicting results and many complications of interpretation, multiple comprehensive reviews conclude that it is difficult to distinguish expectation suppression from the neural adaptation effects that underlie the well-documented repetition suppression effect (Kok & de Lange, 2015;Kok et al, 2012;Lee & Mumford, 2003;Summerfield & Egner, 2009;Vinken & Vogels, 2017;Walsh et al, 2020). Furthermore, detailed single-neuron level recordings are the least likely to show these effects -instead, they are most evident in aggregate signals such as the BOLD response in fMRI, suggesting that they may more strongly reflect population-level differences in activity, rather than individual explicit error coding neurons.…”