The Effect of Working Memory on Corticospinal Excitability

Abstract

Skilled behavior is dependent upon the ability to extract and integrate sensory afference into appropriate output from motor cortex. This process is dynamic with executive control, guided by declarative knowledge (i.e. facts and semantics) and able to shape subconscious processes guided by procedural knowledge. Previous work by Suzuki et al.1 used short-latency afferent inhibition to show that verbal working memory demands (a declarative construct) change afferent projections to the cortical motor output neurons, providing a route by which executive control shapes motor cortical output. Whether other variants of working memory have the same influence on motor output and whether the same neuronal circuits are involved is unknown. Therefore, the current study sought to investigate the influence of spatial working memory on different afferent projections converging on the corticospinal neuron in the motor cortex. Short- (SAI) and long-latency afferent inhibition (LAI) were assessed in seventeen participants during the maintenance period of a spatial or verbal working memory task conducted over the course of two sessions per participant. Either session consisted of one of the two working memory tasks. In the spatial memory task, participants were required to encode a spatial array and maintain the array in working memory to determine whether a probe matched or did not match the original display. The probe consistent of a single dot and participants indicated whether the probe was part of the initial set. The spatial array consisted of either two or six dots around a central fixation cross. In the verbal memory task, participants were required to encode an array of letters and maintain the array in working memory to determine whether a probe matched or did not match the original display. The probe consistent of a single letter and participants indicated whether the probe was part of the initial set. The verbal set consisted of either two or six letters. The effect on different afferent circuits was assessed by manipulating the direction of induced current used in the assessment of SAI and LAI. The “PA” afferent circuit was recruited using TMS induced current in the posterior-anterior direction, and the “AP” afferent circuit was recruited using TMS induced current in the anterior-posterior direction. The order of task and current direction was randomized across participants with TMS current in each direction (AP/PA) conducted within both sessions. Baseline assessments of spatial and verbal working memory capacity evaluated the influence of working memory on sensorimotor circuits. iv Increasing verbal working memory load increased SAI from circuits recruited by current in the PA but not AP direction. Verbal working memory load had no impact on either PA or AP circuits mediating LAI. In contrast, spatial working memory load had no effect on either PA or AP circuits mediating SAI. Instead, increasing spatial working memory load increased LAI recruited by AP circuits. These results suggest that spatial and verbal working memory influence the AP- and PA-mediated afferent circuits that converge on the corticospinal neuron to shape motor output. These different inputs may provide distinct pathways by which declarative knowledge can shape representations of motor skills.