Multiple sclerosis (MS) is an immuno-mediated disorder of the central nervous system (CNS) in which inflammation, demyelination, and axonal damage coexist (1). MS prevalence and incidence have increased progressively over time (2). Importantly, MS is commonly diagnosed between 20 and 40 years of age, affecting the early stages of working lives and leading to a further increase in social costs (3, 4). MS is characterized by sensory and motor symptoms, bladder and bowel problems, cognitive impairment and fatigue. Fatigue trait in MS was defined, by the Multiple sclerosis council for clinical practice guidelines, as “a subjective lack of physical and/or mental energy that is perceived by the individual or caregiver to interfere with usual and desired activities”(5). It occurs frequently and is considered one of the most disabling symptom since it interferes with the performance of daily activities (6). However, the pathophysiology of fatigue and the mechanisms involved are still poorly understood (7–13).

On the other hand, fatigue state, can be defined as the decline in peak force (torque) after performing an exercise intervention (14). Generally, fatigue triggered by overstrain in healthy subjects, is mediated by muscular factors even if changes in the peripheral and central nervous system as well as lack of self-motivation (15) or a subjective decrease in mental/physical energy (16) are also involved (7, 17–19).

In this line, in healthy subjects it was demonstrated that the development of fatigue depends on the changes affecting several structures such as spinal cord, cerebral cortex and subcortical structures (17, 20, 21). Accordingly, it was recently shown that fatigue in MS patients depends on the changes of functionality of CNS (10, 11, 21–26)]. Since the impact of fatigue in daily life of MS patients is high and since 80–90% of MS patients present with fatigue (26), a more profound knowledge of the mechanisms involved in the fatigue development could help to better evaluate patients and ultimately could help to find new treatments.

Although several works seem to suggest that MS fatigue is associated to corticospinal dysfunction, other works did not confirm this causal link (10, 21, 23). Moreover, in healthy subjects not-fatiguing tasks did not lead to reduction in Motor Evoked Potentials (MEP) (i.e., post exercise depression) (27), while studies evaluating post-exercise depression of MEP amplitude in MS patients demonstrated variable results (increased in one study, reduced in another and similar to healthy subjects in 3 studies) and that the amount of post-exercise depression was not related to self-reported fatigue in MS (28).

In the present study, we aimed to give more mechanistic insights about the lack of post exercise depression in fatigued MS patients, by studying the effects of a fatiguing hand task on motor cortex and peripheral nerve functions and by looking for any correlation between neurophysiological parameters of central and peripheral fatigue, maximal handgrip strength (MHS) and subjective perception of fatigue. We hypothesized that the lack of post exercise depression in fatigued MS patients could be explained by different neurophysiological alterations in motor cortex and peripheral nervous system functions.

To study fatigue state, we used a handgrip task of 2 min in which subjects were asked to squeeze a hand dynamometer to produce maximal voluntary contraction: subjects were asked to maintain their maximal strength during the whole duration of the task and were verbally encouraged to provide maximum contractions. We a priori decided to evaluate the left hand in control group while both hands, in two separated sessions, were evaluated in MS patients.

Handgrip task was performed by using a hand-held dynamometer (Cibertec SA, Spain) while the subjects were seated with their shoulder adducted and neutrally rotated, elbow flexed at 90°, forearm in neutral position, and wrist between 0 and 30° dorsiflexion and between 0 and 15° ulnar deviation. During grasping, subjects were instructed to flex their fingers. Data were recorded on computer for later analysis using a CED 1401 A/D converter and Spike software (Cambridge Electronic Design, Cambridge, UK). Maximal handgrip strength (MHS) was defined as the highest peak torque (Newton, N) obtained during the 2 min handgrip task for each participant. The entire dynamometer signal (2 min) was analyzed by dividing it in 20 s' blocks (6 blocks in total, namely B1-B6). For each block, the modulus of the handgrip strength (HS) was quantified and data were expressed as Newton per seconds (N^*s). Force decay (i.e., modulus decay; FD) over the 2 min was considered as a marker of fatigue trait (last 20 s/first 20 s^*100). Modulus is defined as area over the zero baseline but with all absolute values.

At the end of each session, all subjects were asked to report their level of perceived intensity during handgrip using the Borg Rating of Perceived Exertion Scale (BRPES, range 6–20, where 6 means no exertion at all and 20 means maximal exertion). Although this is a subjective measure, a person's exertion rating may provide a fairly good estimate of physical activity (30). The total HS modulus of the whole task (THS- N^*s) was also calculated to quantify the total effort of each subject/patient. The ratio between the whole task modulus (root squared to reduce variability) and BRPES were also calculated. This indicates the effort perception relative to the continuous motor activity.

We aimed to study the effects of an isometric handgrip-fatiguing task on peripheral and central motor excitability of controls and MS participants (case-control study design).

For this reason we used Transcranial Magnetic Stimulation (TMS) to evaluate central motor excitability and peripheral nerve stimulation for peripheral motor excitability.

TMS was performed to evaluate central motor excitability by assessing MEP modulus and central motor conduction time (CMCT) (see below for details) of both upper limbs in MS patients and of left upper limb in controls; these parameters were used as putative markers of corticospinal integrity (31).

Peripheral nerve electrical stimulation (pES) of the ulnar nerve was used to evaluate peripheral motor excitability by assessing compound muscle action potentials (CMAP) of the first dorsal interosseous (FDI) muscle (amplitude and latency) and the F wave latency (see below for details), again for both hands in MS patients and for left hand in controls.

Central motor excitability and peripheral nerve excitability were evaluated before and after a 2-min handgrip task both in MS patients and healthy controls.

Before entering ANOVA data were checked for normality by means of Kolmogorov-Smirnov test. Data presented are uncorrected for multiple comparisons.

All the participants tolerated the whole experimental procedure and none experienced TMS related side effects. One MS participant had no recordable MEPs in the left FDI muscle and this side was excluded from final analysis of the neurophysiological variables. Kolmogorov-Smirnov test showed that data had normal distribution.

Data are summarized in Table 2.

Our data confirmed the expected baseline differences between MS patients and controls: particularly we found a reduced CMAP and MEP modulus and an increased RMT in MS patients when compared to controls. Our data fits well with those of previous studies demonstrating an impairment of corticospinal tract in MS (35, 36, 38, 39). On the other hand, we found a reduction of baseline CMAP modulus in MS patients. Although MS affects CNS, the existence of a subclinical involvement of the peripheral nervous system, thought to be associated to MS or induced by muscular non-use or disability in general, was previously demonstrated (40). In this scenario, we cannot exclude that MEP modulus reduction depends on the subclinical impairment of peripheral nerve targeting FDI. On the other hand, the absence of an impairment of the latency excludes gross peripheral conduction abnormalities and severe axonal damage in the studied peripheral nerve (see also study limitations). Furthermore, the reduction of MEP modulus in MS patients could be drug-dependent since the patients included in the present study were taking several CNS active drugs, as shown in Table 1. However, when the pharmacological treatments of our patients was reviewed, it could be shown that 2 patients were taking clonazepam (a GABA-A agonist), 5 patients were taking baclofen (a GABA-B agonist), 2 patients were taking pregabaline, 1 patient gabapentine, 2 patients amantadine, 1 patient modafinil, 3 patients cannabidiol, and only 4 patients were taking anti-depressant drugs (1 Venlafaxine 1 Fluoxetine and 2 Escitalopram). In this context, just GABA-A agonist could affect motor threshold (i.e., increasing it) and MEP amplitude (i.e., reducing it) whereas the other mentioned drugs did not substantially reduce motor cortex excitability as reported by Ziemann (41). When we take into account these observations we can suppose that the reduction of motor cortex excitability in the MS patients group was not primarily due to a pharmacological effect.

MS patients showed clinically significant fatigue in 2/3 of the patients (11/17), even considering that a few patients were receiving anti-fatigue drugs. Only four patients were taking anti-depressant drugs and none of the patients or subjects referred significant mood disorders during baseline clinical interview. It is demonstrated that mood as well as sleep disorders are common in MS and could participate in the perception of fatigue in daily live activities (42). We did not formally screened for sleep disorders, which in any case were not reported by the participants in the general interview.

As expected, patients presented diminished maximal handgrip strength and corticospinal dysfunction (prolonged CMCT). Moreover, there was a tendency to negative correlation between these latter parameters (the longer the CMCT the less the maximal handgrip strength). On the other hand, fatigue state shown by the motor decay that subjects experienced during the isometric fatiguing task was similar in both MS patients and healthy controls (of course MS patients started the task with much less strength). Both patients and controls had also similar perceived fatigue state of the executed task, but whilst in controls there was a correlation between the force decay and both for general fatigue trait (scores FSS and FR) and fatigue state (Borg scale), in the MS group this correlation was not present. In other words, fatigue trait and fatigue state in MS patients are independent from handgrip strength as well as fatigue state is independent from fatigue trait. Moreover, FSS and force decay were independent from the corticospinal tract dysfunction (CMCT). As FSS concerns, we have to take into account that is a multidimensional scale used for a gross evaluation of fatigue in daily life, so its relationship with the corticospinal tract can be tiny. Furthermore, lack of correlation between corticospinal tract dysfunction and force decay suggests that force decay induced by an isometric task does not depend on corticospinal dysfunction in MS. Differently from control group, we were unable to detect clear MF (by TMS after the handgrip task) or a clear reduction in peripheral nerve excitability in MS. Previous studies reported a lower decay in cortical excitability (43) or an increase (i.e., compensatory) in central motor drive in MS patients (23, 44) even if these authors reported also a similar reduction of peripheral nerve excitability in MS and controls. This discrepancy was probably due to the normality of CMCT in MS patients found in previous studies (43), while patients included in our study were characterized by CMCT alteration. This suggests that in our MS patients, the dysfunction in corticospinal connection (i.e., increased CMCT), reduced the torque realized during the task (MHS lower than controls), impeded, together with the subclinical alteration in peripheral nervous system demonstrated in patients, a reduction in peripheral nerve excitability and induced a compensatory increase in motor cortex excitability, that prevented MEP reduction after fatiguing task.

Following correction, taking in consideration the real total effort performed by the subjects (THS/BRPES), we can note that the perceived fatigue is much more in the MS group. Based on these data, a number of considerations could be suggested with proper caution: (1) Fatigue trait does not depend on corticospinal dysfunction in MS; (2) force decay induced by an isometric task does not depend on corticospinal dysfunction in MS; (3) maximal handgrip strength depends on corticospinal dysfunction in MS; (4) Perceived fatigue in MS is higher than in controls (similar perceived fatigue with much less THS); (5) MEP and CMAP changes after 2-min isometric fatiguing task are not able to measure fatigue in MS, probably due to a central motor drive compensation.

MEP amplitude indirectly measures the descending corticospinal drive and, in healthy subjects, fatigue affects this descending activity (32).

In our experiment, healthy subjects showed a reduction in MEP modulus in line with other previous results that report an increase during the fatiguing task and a post-fatigue reduction in MEP amplitude (17). However, as elegantly described by Gandevia, MEP amplitude increased after 2 min of MVC just for the first 30 s and after 60 s began to decrease (20), so it is not surprising that our data demonstrated a MEP reduction in healthy subjects in the first minutes recording block as facilitation prevails just for the first 30 s. The post-exercise facilitation is probably due to the difficulty to keep the muscle fully relaxed after the fatiguing task (20). Our data on EMG background confirm the reduced relaxation immediately after the task. On the other hand, in the first block after the fatiguing task we observed similar values of the EMG background for controls and MS subjects, suggesting that the difference in MEP changes could not be explained by different muscle relaxation capacity between the two groups.

This point could further suggest that controls show a reduction in corticospinal output, after fatiguing task, while MS patients do not. In this context, it could be considered that fatigue state in MS patients also depends on connectivity changes in motor and non-motor basal ganglia, including motivation and reward circuits (13, 45, 46) and could be also due to adaptive changes that take place in MS pathological process along with the whole duration of the illness, related to clinical and subclinical neuropsychiatric changes, as well as depression and cognitive impairment (47, 48). Furthermore, fatigue-induced cortical hyperexcitability was already described in MS patients (23, 44). The two phenomena could indicate a cortical reorganization that might serve to compensate for progressive tissue damage by MS (49) and could contribute to fatigue pathophysiology in multiple sclerosis (i.e., the more the reorganization the more central fatigue).

It is important to note that the present study presents some limitations that could affect result interpretation.

We recruited the MS study participants in a consecutive manner from a patient association, and for this reason the cohort of the MS group was not homogeneous. Patients were taking different disease-modifying drugs and symptomatic medications that may affect our findings. Unfortunately, due to the limited number of patients included in the study, we could not perform further analysis to differentiate the effects of each drug on our results. For example, it should be considered that interferon could affect fatigue itself, directly but also indirectly, by means of a drug-induced worsening of sleep duration, quality, and stability (50). Moreover, we did not formally screen patients for sleep disorders, so that we cannot quantify and evaluate the sleep-induced effects on fatigue.

Another limitation is the lack of further neurophysiological work-up in our study: particularly we did not evaluate other TMS protocols that can provide information about GABA-A and GABA-B circuits or plasticity. However, our main objective was to observe and study the changes in fatigue induced by a handgrip and its relationship with corticospinal tract, so we considered that paired-pulse TMS and protocols to evaluate plasticity were beyond the main objective of the present study.

Furthermore, we did not analyze the neuroradiological findings at the moment of the study (or close by the study). Similarly, we do not have a formal clinical neurophysiological assessment of the peripheral nervous system (motor and sensory nerve conduction studies, F-waves, H-reflexes, and needle EMG). Obviously, the lack of these kinds of information prevented the execution of further correlations' analysis that could have been provided further relevant details about the phenomena observed.

Since MS patients and healthy subjects present important neurophysiological differences in baseline conditions, also fatigue generation mechanisms were different between the two groups. In healthy subjects, isometric fatiguing task induces a central fatigue that involves corticospinal functions as demonstrated by MEP reduction and confirmed by analyzing data normalized to M-wave excluding by this way a possible effect over the peripheral nervous system. In MS patients, the pathogenesis of fatigue seems not driven by the mechanisms directly related to corticospinal function (that characterize fatigue in controls) but seems probably due to other “central abnormalities” upstream to primary motor cortex (11).

This study was carried out in accordance with the recommendations of Toledo Area Ethical Committee. The protocol was approved by the Toledo Area Ethical Committee. All subjects gave written informed consent in accordance with the Declaration of Helsinki.

LM-M, NM-L, MD, MR, JA, and AO conceived of the study and designed the experimental paradigm. VS-L, MT-P, DP-P, and CA-B performed the experiments and analyzed the data. LM-M and AO wrote the manuscript. LM-M, NM-L, MD, MR, and JA provided feedback and edited the manuscript. All authors read and approved the final manuscript.