Abstract
Introduction Measures of cortico-spinal excitability (CSE) in studies with transcranial magnetic stimulation (TMS) are confounded by physical and physiological covariates, such as coil location, tilt, orientation and pre-innervation (Julkunen et al., 2009). Consequently, inter-trial-reliability is significantly reduced (Darling et al., 2006). Here, the core idea was to model CSE estimates independent of potential confounders. We introduce multilinear regression as our tool to partition the influence of confounders such as pre-innervation, typically controlled for by audio-visual feedback. Hypothetically, variability of the corrected data is associated with the experimental conditions devoid of confounding effects. We investigate and validate this suggestion in previous data collected under differential conditions of focused motor attention arguably confounded by pre-innervation (Bathe-Peters et al., DGKN 2011). Methods 8 healthy, brain stimulation-naive volunteers participated in the study. A modified simple reaction time paradigm cued motor attention to the dominant abductor pollicis brevis (APB), first dorsal interosseus (FDI) and abductor digiti minimi (ADM) (CUE 1–3) muscles. Single TMS-pulses were delivered at four latencies (LAT 1–4) relative to the cue onset always over the predefined FDI hot spot. Pre-innervation was quantified prior to each MEP trigger (100–0ms) by the area under the curve in the respective electromyogram (EMG) trace. MEP amplitudes were subjected to stepwise multilinear regression analysis with pre-innervation and interactions of cue, EMG and latency conditions as possible predictors. Results Inclusion of pre-innervation as a predictor yielded a reduction in variance of about 5%. A strong positive effect of motor attention persisted in interaction with stimulation 50ms prior to expected cortical movement execution: beta values ([95% confidence intervals]) were approximately 347.7 ([240.0; 455.5]) for APB, 224.2 ([148.4; 300.0]) for FDI and 288.7 ([160.8; 416.7]) for ADM, always p Conclusions Utilizing a multilinear model we first identified, and secondly corrected for pre-innervation as a potential confounder in measures of CSE. So far, audio-visual feedback has been the typical method for control; now, we offer a novel procedure which quantifies and corrects for confounders based on their predictive validity. Stepwise regression, that involves an iterative algorithm to find the best possible model, suggested pre-innervation as a predictor of MEP size in every single subject. Thus, residual variability can be attributed to the experimental conditions. Therefore, we confirm with additional evidence that the functional characteristics of the primary motor representation are dynamic and predicted by the allocation of motor attention.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call