Predictive coding in the visual cortex: a functional interpretation of some extra-classical receptive-field effects

Abstract

We describe a model of visual processing in which feedback connections from a higher- to a lower-order visual cortical area carry predictions of lower-level neural activities, whereas the feedforward connections carry the residual errors between the predictions and the actual lower-level activities. When exposed to natural images, a hierarchical network of model neurons implementing such a model developed simple-cell-like receptive fields. A subset of neurons responsible for carrying the residual errors showed endstopping and other extra-classical receptive-field effects. These results suggest that rather than being exclusively feedforward phenomena, nonclassical surround effects in the visual cortex may also result from cortico-cortical feedback as a consequence of the visual system using an efficient hierarchical strategy for encoding natural images.

Computational modeling and exploration of contour integration for visual saliency

“…Additionally, it is easy to imagine that the dopamine-like priming mechanism we have hypothesized here not only enhances contours, but may play an integral part in training the system in a similar manner as suggested in Rao and Ballard (1999). For instance, it has been proposed that observed movement of objects trains neurons to recognize contours (Prodöhl et al 2003).…”

“…Our model agrees with these observations since build up of group suppression increases endstop detection and would also create a delay for such detection as suppression builds. This is also similar to the model by Rao and Ballard (1999), which used a predictive feedback suppression mechanism to facilitate end-stop detection.…”

Predictive processing increases intelligibility of acoustically distorted speech: Behavioral and neural correlates

“…Previous studies (Clos et al., 2014; Sohoglu & Davis, 2016; Sohoglu et al., 2012; Tuennerhoff & Noppeney, 2016) have suggested that predictive coding (Friston, 2005; Huang & Rao, 2011; Mumford, 1992; Rao & Ballard, 1999) underlies the instant increase in intelligibility of distorted speech signals when these are presented simultaneously with or immediately after the presentation of the disambiguating stimulus (e.g., a written or intact auditory counterpart of the speech stimulus). The predictive coding framework proposes that information residing in an internal predictive model is fed back from higher‐order cortical areas to lower‐level brain areas whose activity reflects the difference between auditory input and the predictive information, that is, the prediction error signal (Friston, 2005; Huang & Rao, 2011; Mumford, 1992; Rao & Ballard, 1999).…”

“…The predictive coding framework proposes that information residing in an internal predictive model is fed back from higher‐order cortical areas to lower‐level brain areas whose activity reflects the difference between auditory input and the predictive information, that is, the prediction error signal (Friston, 2005; Huang & Rao, 2011; Mumford, 1992; Rao & Ballard, 1999). This error signal is projected to the higher‐order cortical areas through feedforward connections to update the internal model.…”

The contribution of active body movement to visual development in evolutionary robots

“…Recent advances in computational neuroscience have shown that relatively simple models of developmental visual systems are capable of developing qualitatively similar properties to those found in the early stages of visual processing in cats and monkeys (Hancock et al, 1992;Field, 1994;Olshausen and Field, 1996;Rao and Ballard, 1999). However, those models often use images from publicly available databases or photographs taken in a natural environment as visual stimuli, and do not allow the system to freely interact with the environment and choose those sensory events.…”