Ageing increases reliance on sensorimotor prediction through structural and functional differences in frontostriatal circuits

Abstract

The control of voluntary movement changes markedly with age. A critical component of motor control is the integration of sensory information with predictions of the consequences of action, arising from internal models of movement. This leads to sensorimotor attenuation-a reduction in the perceived intensity of sensations from self-generated compared with external actions. Here we show that sensorimotor attenuation occurs in 98% of adults in a population-based cohort (n=325; 18-88 years; the Cambridge Centre for Ageing and Neuroscience). Importantly, attenuation increases with age, in proportion to reduced sensory sensitivity. This effect is associated with differences in the structure and functional connectivity of the pre-supplementary motor area (pre-SMA), assessed with magnetic resonance imaging. The results suggest that ageing alters the balance between the sensorium and predictive models, mediated by the pre-SMA and its connectivity in frontostriatal circuits. This shift may contribute to the motor and cognitive changes observed with age.

Figure 1. The Force Matching Task.

(a) Schematic representation of the behavioural task, in which participants matched a force applied to their left index finger via a lever attached to a torque motor. In the Direct condition, participants used the right index finger to apply the force directly on the lever, whereas in the Slider condition, they matched the force by moving a linear potentiometer that controlled the torque motor. The main measure for each subject was the mean overcompensation, calculated as the difference between the matched force and target force across the different force levels. (b) Standard boxplots showing the distribution of force overcompensation values across participants in the Direct and Slider conditions (black line indicating the median). A positive value indicates sensory attenuation. Significance of comparisons is indicated by **=P<0.01, ***=P<0.001.

Figure 2. Age-related differences in sensorimotor attenuation.

(a) Mean force overcompensation in the Direct condition (shades of cyan; left panel) and Slider condition (shades of magenta; right panel) across three age groups (error bars indicate standard error of the group mean). Significance of pair-wise comparisons is indicated by *=P<0.05, ***=P<0.001. (b) Overcompensation plotted against age across all participants for Direct condition (left) and Slider condition (right). For illustration, a moving average (window size of 50) is displayed.

Figure 3. Matched force versus target force.

Mean linear regression fits of the matched force against the target force for the Direct (cyan) and Slider (magenta) conditions, plotted separately for ‘young adults' (triangles) and ‘older adults' (squares). These two groups refer to the age groups described in Fig. 2a and in the main text. Mean (±2 standard error of group mean) of matched forces are shown across the different target forces for each age group. Analyses on the slope and intercept of linear regressions were performed with age as a continuous variable (see text), but are illustrated as a categorical effect for simplicity.

Figure 4. Differences in grey matter volume in relation to sensorimotor attenuation.

(a) Results of a Voxel-Based Morphometry analysis, examining the correlation between grey matter volume and Direct force overcompensation. Grey matter volume in the pre-SMA (peak voxel in x=+14, y=+12, z=+58) was negatively associated with increased sensorimotor attenuation in the Direct condition (cyan; P<0.05, FWE-corrected with cluster-forming threshold of P<0.001, uncorrected). Cluster size=1,193 voxels; voxel size=3.375 mm³. (b) The z-score of grey matter in the pre-SMA cluster from a divided by total intracranial volume, plotted as a function of age. The pre-SMA cluster also demonstrated an age-related difference in grey matter volume with (n=302, r=−0.25, P=1.2e−05) and without (n=302, r=−0.29, P=2.1e−07) adjusting for total intracranial volume. For illustration, a moving average (window size of 50) is displayed.

Figure 5. Differences in functional connectivity with pre-SMA in relation to ageing and increased Direct force overcompensation.

(a) Clusters showing increased functional connectivity with the pre-SMA seed from the VBM analysis in relation to ageing and Direct force overcompensation, during a movement task (green), during task-free resting state (red) or both (yellow). Significant (P<0.05, FWE-corrected) clusters in each of the two fMRI datasets were identified at a cluster forming threshold of P<0.001, uncorrected. (b) Same as in a, but for clusters showing reduced connectivity with pre-SMA in relation to ageing and increased Direct overcompensation.