Determinants of Inter-Individual Variability in Corticomotor Excitability Induced by Paired Associative Stimulation

Abstract

Transcranial magnetic stimulation (TMS) is a well-established tool in probing cortical plasticity in vivo . Changes in corticomotor excitability can be induced using paired associative stimulation (PAS) protocol, in which TMS over the primary motor cortex is conditioned with an electrical peripheral nerve stimulation of the contralateral hand. PAS with an inter-stimulus interval of 25 ms induces long-term potentiation (LTP)-like effects in cortical excitability. However, the response to a PAS protocol tends to vary substantially across individuals. In this study, we used univariate and multivariate data-driven methods to investigate various previously proposed determinants of inter-individual variability in PAS efficacy, such as demographic, cognitive, clinical, neurophysiological, and neuroimaging measures. Forty-one right-handed participants, comprising 22 patients with amnestic mild cognitive impairment (MCI) and 19 healthy controls (HC), underwent the PAS protocol. Prior to stimulation, demographic, genetic, clinical, as well as structural and resting-state functional MRI data were acquired. The two groups did not differ in any of the variables, except by global cognitive status. Univariate analysis showed that only 61% of all participants were classified as PAS responders, irrespective of group membership. Higher PAS response was associated with lower TMS intensity and with higher resting-state connectivity within the sensorimotor network, but only in responders, as opposed to non-responders. We also found an overall positive correlation between PAS response and structural connectivity within the corticospinal tract, which did not differ between groups. A multivariate random forest (RF) model identified age, gender, education, IQ, global cognitive status, sleep quality, alertness, TMS intensity, genetic factors, and neuroimaging measures (functional and structural connectivity, gray matter (GM) volume, and cortical thickness as poor predictors of PAS response. The model resulted in low accuracy of the RF classifier (58%; 95% CI: 42 − 74%), with a higher relative importance of brain connectivity measures compared to the other variables. We conclude that PAS variability in our sample was not well explained by factors known to influence PAS efficacy, emphasizing the need for future replication studies.

Primary motor cortex excitability as a marker of plasticity in a stimulation protocol combining action observation and kinesthetic illusion of movement

“…All these studies attempted to find a proxy anticipating the efficacy the treatment would have on the participant. Examples of individual neurophysiological signatures that were found to correlate with the long‐lasting efficacy of the previously mentioned stimulation paradigms are individual motor cortical excitability (operationalized as TMS stimulation level over M1 required to produce MEPs of 1 mV or as resting motor threshold) (Labruna et al., 2019; López‐Alonso et al., 2014; Müller‐Dahlhaus et al., 2008), MEP latency (Wiethoff et al., 2014), strength of short‐interval intracortical inhibition (Murase et al., 2015), level of short‐latency afferent inhibition (Cash et al., 2016), structural connectivity in the corticospinal tract (Minkova et al., 2019) and electroencephalographic response (Iscan et al., 2016).…”

“…Being a “responder” to a stimulation protocol means being a person in whom the protocol was efficient in evoking the behavioral and neurophysiological effects for which it was administered (Huang et al., 2017). Recent studies have dealt with this matter showing high inter‐ and intra‐subject variability in response to different stimulation paradigms known to evoke plasticity in M1; for instance, transcranial direct current stimulation (tDCS) (López‐Alonso et al., 2014; Strube et al., 2015; Wiethoff et al., 2014), paired associative stimulation (Minkova et al., 2019; Müller‐Dahlhaus et al., 2008; Murase et al., 2015; Sale et al., 2007), and theta‐burst stimulation (Rocchi et al., 2018; Vallence et al., 2015) protocols. All these studies attempted to find a proxy anticipating the efficacy the treatment would have on the participant.…”

“…Whether most of the previous studies found correlations between the protocol efficacy and individual neurophysiological parameters that were not directly related to the plasticity‐inducing protocol, the present findings reveal that a neurophysiological response specific to the proposed intervention, namely the increase in motor cortical excitability during AO‐KI plays a crucial role in predicting AO‐KI effectiveness. This specificity is particularly relevant because considering M1 excitability during protocol administration as a proxy of the following plasticity might be (a) easier and faster than other methods involving neuroimaging or electroencephalography techniques (Iscan et al., 2016; Minkova et al., 2019), and (b) more stable than the unspecific measures of M1 excitability and related measurements (Wassermann, 2002).…”

Spontaneous Fluctuations in Oscillatory Brain State Cause Differences in Transcranial Magnetic Stimulation Effects Within and Between Individuals

“…Originally, PAS involved peripheral nerve stimulation that was paired with single-pulse TMS to primary motor cortex in order to enhance corticomotor excitability ( Stefan et al, 2000 ), but PAS has also been employed to facilitate communication between the motor cortex and interconnected cortical areas ( Veniero et al, 2013 ). As for the other plasticity-inducing TMS protocols, there is high inter-subject variability in the effects of PAS ( Sale et al, 2007 ; Florian et al, 2008 ; López-Alonso et al, 2014 ), with a recent study reporting that only 61% of participants responded to PAS ( Minkova et al, 2019 ).…”

Convergent Associative Motor Cortical Plasticity Induced by Conditional Somatosensory and Motor Reaction Afferents

“…This study investigates whether intentional, active movements driven by the contralateral M1 influence the M1 plasticity induced by the conventional PAS protocol. It has been reported that the conventional PAS protocol has a responsive variability (Lahr et al, 2016;Minkova et al, 2019). Since the interval between the median nerve stimulation and the TMS is fixed to 25 milliseconds (ms) in the conventional PAS, one of the possible factors causing individual variability would be dispersed conditioning stimuli at M1 corresponding to the individual timing of the somatosensory afferent.…”