The automatic brain: studies on practice and brain function in healthy subjects and patients with schizophrenia

Abstract

Practice makes perfect. The neural mechanisms behind the behavioral improvement of practice (automatization) however are largely unknown. Here we investigate how practice changes brain function and how this can improve our processing capacity. We also examine whether a deficit in automatization can explain the severely limited processing capacity in schizophrenia. Previous research implicates working memory (WM) in the development of automatization and the ability to improve processing capacity with practice. The course of activity changes in brain regions important for WM following practice for distinct behavioral components however is poorly understood. We show that practice induces heterogeneous changes in WM activation for information encoding and responding to information. Practice predominantly facilitates information encoding and thereby possibly improves the capacity to process otherwise interfering information. Another question pertains to whether decreases in activity in brain areas important for WM are accompanied by compensatory changes elsewhere in the brain. We show that practice neither leads to a shift from one network to another, nor to a gain in involvement of regions within the initially active network(s). Practice thus induces a general decline in the WM system without compensatory changes elsewhere in the brain. This may suggest that WM allows restructuring of information that allows more efficient processing following practice as encoding strategies performed by WM decrease. By selectively interfering with activity in brain regions important for WM we investigated the critical importance of WM for automatization. We show that automatization diminishes critical contributions of WM to performance and that with sufficient practice, performance may become independent of cognitive control. This supports the idea that WM engages in restructuring of information when performance is novel, thereby enabling more efficient processing as WM slowly disengages over the course of practice. Patients with schizophrenia typically exhibit inefficient brain function when task load is within the boundaries of their capacity. We investigated whether a failure in automatization can explain inefficient WM function and reduced capacity in schizophrenia. Although patients showed increased (and thus inefficient) brain activity during novel performance, automatization was not impaired. In addition, there was no clear relationship between automatization and the deficit in patients to concurrently perform two tasks. The results suggest that inefficient WM function and reduced capacity in schizophrenia are associated with a failure to properly engage WM when task demands are high for instance during novel performance and when performing an additional task concurrently when goals require frequent updating. This research shows that automatization involves a dynamic distribution of processing resources that allows organizing and structuring of the large amounts of complex information present in our environment. This enables more efficient processing, thereby possibly increasing the capacity to process otherwise interfering information. Severely limited processing capacity in schizophrenia however was not explained by a deficit in automatization. Here we postulate that due to an inability to process frequently changing information, patients with schizophrenia will tend to engage in automatic behaviors, resulting in limited capacity to process information that require flexible and adaptive cognitive strategies.