New frontiers in stroke rehabilitation aim to improve functional recovery taking advantage from the knowledge of mechanisms of cortical reorganization that occur after the acute event (Schaechter, 2004), the so-called “top-down” approach (Chisari, 2015). Literature findings have provided insight into the mechanisms of behavioral rehabilitation techniques, such as constraint-induced movement therapy (Liepert, 2006), and have led to the development of cortical stimulation protocols to improve upper limb recovery (Ward and Cohen, 2004; Nowak et al., 2008; Chisari et al., 2014). The main hypothesis is that an imbalanced inter-hemispheric inhibition occurs following stroke, so the purpose of various rehabilitation approaches is to increase excitability of perilesional intact regions of the affected hemisphere and/or to decrease excitability of the contralesional hemisphere (Hummel and Cohen, 2006; Nowak et al., 2009). Anyway, until now still no customized treatment has been proposed strictly based on the correlations between neurophysiological and functional evaluations.

Transcranial magnetic stimulation (TMS) is a valid tool to obtain data about cortical reorganization (Rossini et al., 2003). TMS is currently used to elicit motor evoked potential (MEP), recorded by surface electromyography (EMG). MEP presence, as a measure of cortical excitability changes and corticospinal tract integrity, offers useful prognostic information about functional outcome (Hendricks et al., 2002; Brouwer and Schryburt-Brown, 2006; Pizzi et al., 2009; Stinear et al., 2014). In pre-activated muscles, TMS may also induce a transient suppression of the EMG-activity after MEP, the so-called silent period (SP) (Kukowski and Haug, 1991; Uozumi et al., 1992), as an inhibitory effect. SP is reported to be abnormally increased in the paretic hand after a stroke (Haug and Kukowski, 1994; Braune and Fritz, 1996; Harris-Love et al., 2016) and tend to decrease with motor recovery. To date, studies on the role of the SP in predicting motor recovery after severe stroke showed rather inconsistent results (van Kuijk et al., 2005, 2014).

Starting from the study conducted by Liepert et al. (2005) the impact of lesion location on motor excitability and motor performance was investigated. The authors evaluated patients with pure motor strokes in four different brain areas (motor cortex lesions, striatocapsular lesions, lacunar lesions of the internal capsule and paramedian pontine lesions), concluding that lesion location determines a specific pattern of motor excitability changes. Recently Thickbroom et al. (2015) highlighted that both the anatomical level of the lesion and to the degree of paretic motor impairment are related cortical excitability and reorganization after stroke. These findings suggest that rehabilitative trials should stratify patients basing on lesion type.

Coupar et al. (2011) also suggest that integrating early clinical data with neurophysiological measurements could be useful to predict long-term recovery and outcome. A remarkable example is the study conducted by Di Lazzaro et al. (2010), which evaluated whether long-term potentiation (LTP)- and long-term depression (LTD)-like changes produced by intermittent theta burst stimulation (iTBS) in acute stroke correlate with outcome at 6 months. They recruited ischemic stroke patients (both with cortical and subcortical lesions) in the first 10 days after stroke, finding that functional recovery is directly correlated with LTP-like changes in affected hemisphere (AH) and LTD-like changes in unaffected hemisphere (UH) and inversely correlated with the baseline excitability of UH. Nevertheless, it is suitable to underline that neurophysiological data during early period post-stroke may suffer from wide inter-subject variability. In particular, Swayne et al. (2008) found that day-to-day variation in clinical performance was unrelated to physiological measures in the first days after the acute event.

Following these considerations our hypothesis was that different stroke lesion location may imply differences in the mechanisms of brain reorganization that lead motor recovery in subacute phase. Our aim was to identify neurophysiological parameters that can be used as markers to describe motor recovery and as factors to guide neurorehabilitation treatment in subacute stroke patients with different lesions.

For this reason we correlated neurophysiological and functional features in a cohort of stroke patients recruited in a specific time window from the acute event and subdivided in cortico-subcortical and subcortical strokes; evaluations were also performed at 3 months after stroke to monitor changes in brain reorganization and clinical behavior.

Our main finding is that the group of patients with subcortical stroke evaluated in this study presented a reduction in contralateral Silent Period (SP) duration in the unaffected hemisphere and this trend is related to clinical improvement 3 months after the acute event. It is well known that the SP can be divided in two parts: an early part principally due to spinal mechanisms and the late part due to inhibition mechanisms of the motor cortex (Chen et al., 1999) mediated through type B-GABAergic interneurons receptors (Di Pino et al., 2014). Literature reported that SP is prolonged in the acute phase after an injury (Liepert et al., 2000b) and it normalizes with the clinical improvement (Classen et al., 1997). In this study we observed a no significant reduction of cSP duration in the AH at 3 months, but this trend (Δ cSP) correlated with functional recovery. Otherwise, the reduction of unaffected intracortical inhibition in S patients may play an important role in effective motor recovery at 3 months.

Changes in cortical activity after stroke have already been investigated in recent literature. Previous studies (Liepert et al., 2000a; Manganotti et al., 2002; Swayne et al., 2008) reported that RMT values were higher on the AH than on the UH. However, the baseline excitability of the AH seems to be less relevant: Swayne et al. (2008) showed a poor correlation between AH neurophysiological features and clinical behavior at 6 months after stroke. The main hypothesis was that the reorganization of alternative cortical networks was more relevant for motor recovery than the function of the original corticospinal tract spared by the ischemic lesions. The role of UH motor cortex in post-lesional recovery is still debated (Takechi et al., 2014). It was supposed that unaffected hemisphere facilitates control of recovered motor function by operating at a higher-order processing level involved in selection, preparation, temporal or spatial organization of movement (Gerloff et al., 2006). Bütefisch et al. (2003) compared stroke patients with good recovery of hand function and controls to determine changes in the inhibitory and excitatory activity in UH motor cortex in stroke patients. They found that in patients' contralesional motor cortex the balance of excitatory and inhibitory activity was shifted toward an increase of excitatory activity; they postulated that in stroke subjects this issue may be relevant for reorganizational processes and may play a role in the functional recovery of patients.

The present study showed substantial stability of RMT values in UH in S group, whereas CS patients showed a no significant increase in RMT value over time. In literature, Manganotti et al. (2002) reported a decrease of RMT in UH detectable in a very early phase after the acute event (5–7 days after stroke) and a trend toward the normalization in patients who experimented a good recovery. Contrarily, the persistence of abnormal excitability in UH was associated with a poor motor recovery. Our data highlighted a substantial stability of UH excitability within 3 months in subcortical patients. The different timing of neurophysiological evaluations may have contributed to mask the effect due to a wider variability. In fact, we described the UH excitability trend in a window between 10 and 45 days after stroke, whereas Manganotti et al. (2002) chose an earlier timing. Swayne et al. (2008) affirmed that the earliest period after stroke is characterized by a wide variability in cortical reorganization and this could account for the different results of our work. Probably, a larger and more homogeneous dataset would help us to resolve this issue and establish a proper timing to characterize neurophysiological changes and correlations with clinical outcome.

Moreover, as already described in literature the presence of MEP in the earliest phase after stroke has a positive predictive value for motor recovery (Stinear et al., 2014). TMS responses were not collected before the first 10 days, so MEP presence at baseline cannot be considered as a predictive factor in our study. However, it emerged that patients with MEP in the affected side at t0 had a better basal clinical score, but not a greater degree of motor recovery.

Regarding the meaning of post-lesional changes observed in the unaffected motor cortex, Shimizu et al. (2002) interpreted the post-lesional cortical changes in the contralateral side as “motor cortex disinhibition,” i.e., a consequent manifestation of the lesion itself. Our data indicate that changes in UH motor cortex could be well described with the modifications detectable with cSP. Moreover, in S patients the cSP demonstrated a clear-cut correlation with clinical outcome.

In conclusion, our analysis proved that cSP is a valid parameter to characterize cortical reorganization patterns in subacute stroke subjects. Furthermore, this measure significantly correlated with motor function in S patients and provided useful information about motor recovery within 3 months.

GL recruited patients, performed clinical and neurophysiological evaluations, contributed to data analysis, interpreted the results and drafted the manuscript; CF performed neurophysiological evaluations, executed data analysis and statistics, helped interpret the results; GS performed clinical and neurophysiological evaluations, contributed to data analysis, helped draft the manuscript; FB performed neurophysiological evaluations, helped to execute data analysis and interpret the results; AP helped perform data analysis and statistics and revised the manuscript; SM was responsible for data analysis and statistics; BR supervised the manuscript; CC designed the study, was responsible for patients' enrolment, contributed to analysis and interpretation of data and revised manuscript. All authors read and approved the final manuscript.

The authors disclose receipt of the following financial support for the research, authorship, and/or publication of this article: this work was supported by the Fondazione Pisa (GRANT: 158/2011). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.