This blog is part of a series on the implications of predictive processing for clinical practice. 

Several people have asked us about praxis and predictive processing and we thought it was useful to write a separate blog on this. We will start by discussing 2 models of sensory processing.

1.       Classic detector model of sensory processing. This has been the model for years.

2.       Predictive processing model of sensory processing. This is now the leading model of sensory processing in neuroscience literature but is less well known in the clinical area.

1.Classic detector model of sensory processing

In the classic detector model, sensory input (bottom up)  is detected,  processed and integrated to produce actions/motor input. Some sensory input becomes habituated and therefore ‘tuned out’. Some actions are based on the integration of the sensory input and some actions are also automatic (top down). For example to swing a bat, the eyes may register the ball, the position of the arm , integrate this information, make a motor plan that results in action. This can then become automated.

2 Predictive processing model of sensory processing

In the predictive processing model, sensory input is predicted from a top down unconscious model and it is only if sensory input is unexpected that the prediction model is then updated. The unexpected sensory input is known as a prediction error. Despite its name, the term prediction error does not mean that it is wrong or incorrect an error, just that it was not predicted by the prior model. updating the model. A further consideration to note is that the brain may decide to update the model or disregard the input, depending on a few factors that are described as  weighting.

In the predictive process model, the brain is trying to minimise prediction errors and to do this , the body can move to make a prediction come true. For example if the brain predicts you will swing a bat at a ball, then the body will do this and then the prediction comes true and remains unaltered. This would occur ‘automatically’ and without conscious awareness. (top down) The only sensory input that would be processed would be unexpected sensory input. So if the swinging the bat at the ball proceeds as predicted, then the sensory input would not need to be registered or processed.

Another example is that if we were walking down stairs, our brain would have a usual ( prior) model of the coordinated body movement and so that would be automatic. If the step was wobbly (and hadn’t been before), this would be unexpected and would be a prediction error. We would have sensory input transmitted to the brain with information on the step, and necessary body movements and the usual (prior) model would be updated.

Interventions

One of the things to keep in mind is that predictive processing is a theory of sensory processing,  and that interventions that work, still work, it is just that the theory behind them may be slightly different and there may be a change of emphasis within the interventions.

For Autistic children and by extension other people who also have sensory processing difficulties,  there is an increased number of ‘prediction errors’ due to not finding the actions predictable and not generalising the actions. Eg seeing that some actions are more unique than they are. This increases the amount/load of sensory input which can then lead to sensory overload if there is too much of it. It may also be useful to practice body movements in the context that they are being used repetitively so that the brain can learn the prediction.

Further information on predictive processing is available in our course: https://sensorymodulationbrisbane.ticketspice.com/predictive-processing-mental-health-and-sensory-processing