Reduction in short interval intracortical inhibition from the early stage reflects the pathophysiology in amyotrophic lateral sclerosis: A meta‐analysis study

Abstract

Background and purpose

Cortical hyperexcitability has been identified as a diagnostic and pathogenic biomarker of amyotrophic lateral sclerosis (ALS). Cortical excitability is assessed by transcranial magnetic stimulation (TMS), a non‐invasive neurophysiological technique. The TMS biomarkers exhibiting highest sensitivity for cortical hyperexcitability in ALS remain to be elucidated. A meta‐analysis was performed to determine the TMS biomarkers exhibiting the highest sensitivity for cortical hyperexcitability in ALS.

Methods

A systematic literature review was conducted of all relevant studies published in the English language by searching PubMed, MEDLINE, Embase and Scopus electronic databases from 1 January 2006 to 28 February 2023. Inclusion criteria included studies reporting the utility of threshold tracking TMS (serial ascending method) in ALS and controls.

Results

In total, more than 2500 participants, incorporating 1530 ALS patients and 1102 controls (healthy, 907; neuromuscular, 195) were assessed with threshold tracking TMS across 25 studies. Significant reduction of mean short interval intracortical inhibition (interstimulus interval 1–7 ms) exhibited the highest standardized mean difference with moderate heterogeneity (−0.994, 95% confidence interval −1.12 to −0.873, p < 0.001; Q = 38.61, p < 0.05; I ² = 40%). The reduction of cortical silent period duration along with an increase in motor evoked potential amplitude and intracortical facilitation also exhibited significant, albeit smaller, standardized mean differences.

Conclusion

This large meta‐analysis study disclosed that mean short interval intracortical inhibition reduction exhibited the highest sensitivity for cortical hyperexcitability in ALS. Combined findings from this meta‐analysis suggest that research strategies aimed at understanding the cause of inhibitory interneuronal circuit dysfunction could enhance understanding of ALS pathogenesis.

INTRODUCTION

Amyotrophic lateral sclerosis (ALS) is a rapidly progressive and fatal neurodegenerative disorder characterized by progressive upper and lower motor neuron dysfunction [1]. Cortical hyperexcitability has been identified as an early and intrinsic feature of ALS [2, 3, 4], mediated through a combination of cortical disinhibition and an increase in facilitation [5, 6, 7]. At a clinical level, cortical hyperexcitability may manifest as hyper‐reflexia and spasticity [8], potentially accounting for the split‐hand phenomenon as well as concordance between handedness and disease‐onset site [9, 10]. Cortical hyperexcitability has been consistently identified as a pathogenic and diagnostic biomarker of ALS [7], preceding lower motor neuron dysfunction [3, 7] and contributing to disease evolution and adverse prognosis [9, 11].

Transcranial magnetic stimulation (TMS) is a non‐invasive technique for assessing cortical motor function [12]. Paired‐pulse TMS assesses cortical interneuronal function by delivering a modulating subthreshold conditioning stimulus prior to a test stimulus. The original ‘constant‐stimulus’ method measured the reduction in the test stimulus motor evoked potential (MEP) amplitude at an interstimulus interval (ISI) of 1–5 ms, termed short interval intracortical inhibition (SICI), a biomarker of cortical inhibitory interneuronal circuits acting via γ‐aminobutyric acid type A (GABA_A) receptors [12, 13]. Subsequently, a threshold tracking technique was developed to overcome the potential limitation of MEP variability, whereby SICI was heralded by greater conditioned‐test stimulus intensity required to generate and maintain the target MEP response of fixed amplitude (0.2 mV ± 20%) [14, 15]. The reduction of SICI has been consistently identified as a biomarker of cortical hyperexcitability in ALS with diagnostic and pathogenic implications [2, 3, 4, 9, 10, 11, 16, 17, 18]. An increase in intracortical facilitation (ICF), between ISIs 10–30 ms, has also been reported as a biomarker of cortical hyperexcitability in ALS [7]. The combination of reduction in SICI along with an increase in ICF and MEP amplitude as well as reduction in resting motor threshold results in an imbalance between cortical inhibition and facilitation leading to hyperexcitability in ALS [18].

Single‐pulse TMS has yielded additional biomarkers of cortical hyperexcitability in ALS [5, 9]. Specifically, reduction of cortical silent period duration has been reported as a biomarker of cortical hyperexcitability in ALS and reflects dysfunction of long latency GABAergic inhibitory circuits acting via GABA_B receptors [2, 16, 17, 19, 20]. Increased MEP amplitude has also been a well described biomarker of cortical dysfunction in ALS, indicating enhanced corticomotoneuronal glutamatergic activity [3, 7, 10, 16, 17, 20]. Reduction in resting motor threshold (RMT) has been described in ALS, although conflicting results have been reported, with some studies reporting normal [2, 3, 9, 21] or even increased RMTs [22, 23].

Transcranial magnetic stimulation (TMS) derived measures of cortical motor function could be of clinical utility in ALS, potentially serving as a surrogate diagnostic and prognostic biomarker, although it remains to be determined which of these biomarkers is of greatest clinical utility. Whilst reduction of mean SICI (between ISIs 1–7 ms) was reported to be the most sensitive TMS measure in ALS [7], replication in larger patient cohorts is required. Consequently, the aim of this meta‐analysis was to investigate which TMS‐derived biomarker of cortical motor function exhibited the greatest utility in differentiating ALS from controls, and to determine the impact of ALS disease duration, functional status and disease‐onset site on clinical utility.

MATERIALS AND METHODS

Literature search

The first study reporting the clinical utility of threshold tracking TMS in ALS was published in 2006 [3]. Consequently, a systematic literature review was conducted of all relevant studies published in the English language by searching PubMed, MEDLINE, Embase and Scopus electronic databases from 1 January 2006 to 28 February 2023. The search for eligible studies to be included in the meta‐analysis was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta‐Analyses (PRISMA) guidelines [24]. The search strategy comprised keywords and MeSH terms, in isolation or combination, and included ‘amyotrophic lateral sclerosis’, ‘cortical hyperexcitability’, ‘motor neuron disorders’, ‘neurodegenerative disease’, ‘neuromuscular disease’, ‘primary lateral sclerosis’, ‘progressive muscular atrophy’, ‘threshold tracking transcranial magnetic stimulation’ or ‘TMS’. The meta‐analysis was registered in PROSPERO (ID#: CRD42023428881).

Data collection and quality assessment

The inclusion criteria included (i) studies reporting threshold tracking TMS using the serial ascending method [14] in ALS and controls, also referred to as the Sydney method by Tankisi et al. [25]; and (ii) studies reporting at least one threshold tracking TMS parameter, SICI. In addition, the use of a 90‐mm circular coil and recording from the abductor pollicis brevis muscle was an additional criterion.

Exclusion criteria included (i) studies reporting constant stimulus or parallel threshold tracking TMS methods; (ii) case reports, reviews, letters or editorial; (iii) conference abstracts; (iv) small studies (sample size <10); or (v) studies reporting the same ALS patients and/or control cohorts. A concerted effort was undertaken to avoid duplication of subjects (ALS patients and controls) across the included papers, and due to this concern a number of potential papers were excluded from the current meta‐analysis [26, 27, 28, 29, 30, 31, 32, 33]. Additionally, the number of subjects included for each study is depicted in Table 1. Two authors (MH and SV) independently reviewed eligible studies to ensure fulfilment of inclusion and exclusion criteria, and a final list of studies was assembled based on author consensus. The Newcastle–Ottawa Scale was used to examine study quality. Data collection was based on published studies only and consequently informed consent and ethical committee review were not required.

TABLE 1.

Summary of studies included in the meta‐analysis with the number of amyotrophic lateral sclerosis (ALS) patients and controls.

Study	Year	ALS	Controls
Suzuki et al. [34]	2022	53	30
Higashihara et al. [35]	2021	40	42
Agarwal et al. [36]	2021	39	30
van den Bos et al. [37]	2021	17	18
Menon et al. a [38]	2020	345	99
Dharmadasa et al. [11]	2020	138	114
Menon et al. a [39]	2019	60	28
van den Bos et al. [40]	2018	27	25
Menon et al. [9]	2017	50	45
Geevasinga et al. [20]	2017	20	11
Matamala et al. [41]	2017	21	40
Shibuya et al. [42]	2016	169	109
Geevasinga et al. [43]	2016	19	31
Menon et al. [44]	2016	18	60
Grieve et al. [45]	2016	25	30
Geevasinga et al. [17]	2015	15	74
Menon et al. [46]	2015	24	33
Menon et al. a [7]	2015	209	68
Menon et al. [10]	2014	26	21
Geevasinga et al. [47]	2014	82	34
Bae et al. [48]	2014	17	15
Vucic et al. [49]	2013	25	30
Vucic et al. [4]	2008	57	55
Vucic et al. [16]	2007	11	26
Vucic et al. [3]	2006	23	34

^a All controls were ALS mimics, whilst the remainder were healthy controls.

The TMS outcome biomarkers included the difference in mean SICI between ISIs 1–7 ms between ALS patients and controls, as well as differences in mean ICF (ISI 10–30 ms), MEP amplitude (expressed as percentage of compound muscle action potential), maximum cortical silent period (CSP) duration (ms) and RMT (% maximum stimulator output).

Clinical features were also collated and included the ALS Functional Rating Scale Revised (ALSFRS‐R) [50], disease duration (months), site of disease onset and upper motor neuron (UMN) score [51]. The UMN score includes biceps brachii, triceps, supinator, finger, patellar, ankle reflexes and plantar flexor reflexes measured on both sides, along with facial and jaw jerks with the score ranging from 0 (no UMN signs) to a maximum score of 16, signifying severe UMN signs [51].

Statistical analysis

Meta‐analysis

Standardized mean differences (SMDs, Cohen's d), standard error and 95% confidence intervals (CIs) were estimated for TMS parameters recorded for ALS patients and controls. Random‐effects models were used to determine SMD. Heterogeneity was assessed using the Higgins index (I ²). Publication bias was assessed by using a funnel plot, Egger's weighted regression method [52] and Duval and Tweedie's trim and fill method [53].

Meta‐regression

Meta‐regression analyses were performed to confirm the association between TMS measures and the following clinical parameters: (i) duration of disease (months), (ii) ALSFRS‐R, (iii) site of disease onset, (iv) gender and (v) UMN score. Meta‐regression analysis was performed with a random‐effects model using restricted maximum likelihood estimation. Given the exploratory purposes of these analyses, combinations were selected for which the p value of the regression coefficient of the clinical parameters was less than or equal to 0.1 and in keeping with previous meta‐analyses [54, 55]. All other data are expressed as mean ± standard deviation or median (interquartile range), with p values <0.05 considered statistically significant.

RESULTS

Study selection and characteristics

The study selection process is described in the flow diagram (Figure 1). During the initial search, 40 studies were identified from the PubMed, MEDLINE, Embase and Scopus databases. After a screening and eligibility criteria assessment 25 studies met the inclusion criteria. The Newcastle–Ottawa Scale score was ≥7 for all included studies indicating high quality.

FIGURE 1.

PRISMA flow diagram detailing the bibliographical searches carried out, including the number of papers screened and the reasons for exclusion of papers from further analysis.

In total, 1530 ALS patients and 1102 controls (healthy, 907; neuromuscular, 195) were assessed with threshold tracking TMS in the 25 studies. Demographic, clinical and TMS characteristics of all participants are summarized in Table 2. In ALS, limb‐onset disease was evident in 69% and bulbar‐onset disease in 31% of patients. The median ALSFRS‐R score was 42 (40.7–42.8) indicating moderate functional disability, whilst the median UMN score was 12 (10.8–12) being consistent with the presence of prominent UMN signs. Additionally, median disease duration in ALS patients was 15.3 (11–17.5) months, indicating that assessment occurred in the early stages of the disease. Sixteen studies reported on riluzole use (N = 1150), with 691 (60%) ALS patients reportedly receiving treatment at time of TMS testing. Apart from riluzole therapy, the use of medications that could impact on TMS findings was not reported in the included studies.

TABLE 2.

Summary of clinical and transcranial magnetic stimulation (TMS) characteristics of amyotrophic lateral sclerosis (ALS) patients and controls.

	ALS	Controls
Number of participants	1530	1102
Median age at assessment (years)	60.7	53
(IQR)	(59–63)	(48–58)
Males:females	933:601	530:437
Median disease duration (months)	15.3
(IQR)	(11–17.5)
Site of disease onset
Limb	68.6%
Bulbar	31.4%
Median ALSFRS‐R	42
(IQR)	(40.7–42.8)
Median UMN score	12
(IQR)	(10.8–12)
Mean SICI between ISIs 1 and 7 ms (%)	2.7	11.1
(SD)	(8.7)	(6.7)
Mean CSP duration (ms)	179.2	204.8
(SD)	(44.6)	(32.7)
Mean ICF (%)	−2.3	−0.5
(SD)	(6.8)	(5.3)
Mean MEP amplitude (%)	37.5	26.3
(SD)	(26.3)	(17)
Mean RMT (%)	58.3	58.8
(SD)	(11.9)	(9.8)

Note: The average of the mean TMS parameters is reported. The control population comprised healthy controls and patients with neuromuscular mimicking disorders including Kennedys' disease, autoimmune neuropathies (chronic inflammatory demyelinating polyradiculoneuropathy, multifocal motor neuropathy), acquired neuromyotonia, facial onset sensory and motor neuronopathy syndrome, spinal muscular atrophy, Hirayama disease, myasthenia gravis. All data are expressed as mean (standard deviation, SD) or median (interquartile range, IQR).

Abbreviations: ALSFRS‐R, Amyotrophic Lateral Sclerosis Functional Rating Scale Revised; CSP, cortical silent period; ICF, intracortical facilitation; ISI, interstimulus interval 1–7 ms; MEP amplitude, motor evoked potential amplitude expressed as a percentage of the compound muscle action potential amplitude; RMT, resting motor threshold; SICI, short interval intracortical inhibition; UMN, upper motor neuron.

Measures of cortical inhibition

Short interval intracortical inhibition

Reduction in mean SICI, between ISI 1–7 ms, has been reported as a robust biomarker of cortical hyperexcitability in ALS [7]. Consequently, all selected studies reported differences in mean SICI (ISI 1–7 ms) between ALS patients and controls. A significant reduction of mean SICI was reported in ALS patients with evidence of moderate heterogeneity (SMD −0.99, 95% CI −1.12 to −0.87, p < 0.001; Q = 38.61, p < 0.05; I ² = 40%; Figure 2). Significant funnel plot asymmetry was evident when assessed by Egger's test with an estimated intercept value being −0.44 (p = 0.002). Importantly, the trim and fill method indicated that ‘study imputation’ did not exert significant effects on SMD (−0.854, 95% CI −0.987 to −0.772, p < 0.0001), suggesting that publication bias was not a concern.

FIGURE 2.

Forest plot depicting effect sizes for individual studies and an overall effect size for short interval intracortical inhibition (SICI). The dotted line represents an absence of significant difference in SICI between amyotrophic lateral sclerosis patients and controls.

Meta‐regression analysis suggested that disease duration, ALSFRS‐R, site of disease onset, UMN score or gender (proportion of males) were not significant predictors of the relationship between mean SICI and ALS (Table 3).

TABLE 3.

Random effects meta‐regression.

	Disease duration (months)	ALSFRS‐R	Site of disease onset	UMN score	Gender
SICI
Coefficient	−0.008	0.030	0.004	−0.009	−0.794
Standard error	0.008	0.050	0.004	0.031	1.415
95% CILOWER	−0.020	−0.06	−0.010	−0.070	−3.567
95% CIUPPER	0.008	0.130	0.012	0.051	1.980
Z value	−1.01	0.71	0.85	−0.310	−0.56
p value	0.314	0.478	0.396	0.759	0.575
CSP duration
Coefficient	−0.003	0.021	−0.004	−0.004	−1.630
Standard error	0.009	0.057	0.005	0.042	1.715
95% CILOWER	−0.021	−0.090	−0.014	−0.085	−4.991
95% CIUPPER	0.015	0.132	0.007	0.078	1.731
Z value	0.340	0.370	−0.730	−0.090	−0.950
p value	0.734	0.714	0.467	0.931	0.342
ICF
Coefficient	−0.014	0.038	0.006	−0.009	−2.404
Standard error	0.008	0.053	0.005	0.046	1.666
95% CILOWER	−0.030	−0.066	−0.005	−0.098	−5.670
95% CIUPPER	0.003	0.142	0.016	0.081	0.862
Z value	−1.65	0.72	1.06	−0.190	−1.44
p value	0.099	0.471	0.289	0.852	0.149
MEP amplitude
Coefficient	0.005	−0.063	−0.002	−0.020	−1.787
Standard error	0.008	0.051	0.005	0.042	1.612
95% CILOWER	−0.011	−0.163	−0.012	−0.102	−4.948
95% CIUPPER	0.021	0.037	0.007	0.062	1.373
Z value	0.62	−1.23	−0.51	−0.480	−1.11
p value	0.54	0.22	0.61	0.631	0.27
RMT
Coefficient	0.010	−0.088	−0.010	0. 037	1.141
Standard error	0.010	0.063	0.006	0.042	2.176
95% CILOWER	−0.010	−0.211	−0.022	−0.046	−3.124
95% CIUPPER	0.030	0.035	0.002	0.120	5.406
Z value	1.000	−1.400	−1.690	0.870	0.52
p value	0.32	0.16	0.09	0.382	0.60

Note: The impact of disease duration from symptom onset, ALSFRS‐R, site of disease onset (limb vs. bulbar) and gender (male vs. female) on TMS parameters was assessed. Given the exploratory purposes of the meta‐regression analysis, combinations were selected for which the p value of the regression coefficient of the clinical parameters was ≤0.1.

Abbreviations: ALSFRS‐R, Amyotrophic Lateral Sclerosis Functional Rating Scale Revised; CI, confidence interval; CSP, cortical silent period; ICF, intracortical facilitation; MEP amplitude, motor evoked potential amplitude expressed as a percentage of the compound muscle action potential amplitude; RMT, resting motor threshold; SICI, short interval intracortical inhibition; TMS, transcranial magnetic stimulation; UMN, upper motor neuron.

Cortical silent period duration

A reduction in CSP duration has also been previously reported in ALS [7]. Meta‐analysis disclosed that reduced CSP duration was significantly associated with ALS with evidence of heterogeneity (−0.66, 95% CI −0.81 to −0.50, p < 0.001; Q = 54.73, p < 0.001; I ² = 64%; Figure 3). Of relevance, the magnitude of effect size was less for CSP duration compared to the effect size for SICI. There was no significant funnel plot asymmetry (Egger's test estimated intercept −0.05, p = 0.81).

FIGURE 3.

Forest plot depicting effect sizes for individual studies and an overall effect size for cortical silent period (CSP) duration. The dotted line represents an absence of significant difference in CSP duration between amyotrophic lateral sclerosis patients and controls.

Meta‐regression analysis suggested that disease duration, ALSFRS‐R, site of disease onset, UMN score or gender (proportion of males) were not significant predictors of the relationship between mean CSP duration and ALS (Table 3).

Measures of cortical facilitation

Intracortical facilitation

An increase in ICF (between ISIs 10 and 30 ms) has been previously reported in ALS patients, being represented by a greater mean negative ICF value [7]. Meta‐analysis revealed a significant increase of ICF in ALS patients compared to controls (−0.24, 95% CI −0.39 to −0.10, p < 0.001; Q = 44.35, p < 0.001; I ² = 57%; Figure 4), with evidence of heterogeneity. There was no significant funnel plot asymmetry as assessed by Egger's test with an estimated intercept value being 0.35 (p = 0.06).

FIGURE 4.

Forest plot depicting effect sizes for individual studies and an overall effect size for intracortical facilitation (ICF). The dotted line represents an absence of significant difference in ICF between amyotrophic lateral sclerosis patients and controls.

Meta‐regression analysis disclosed that disease duration was potentially related to mean ICF changes in ALS (p = 0.1, Table 3). Consequently, a subgroup analysis was undertaken which disclosed a trend for ICF increase in ALS patients with longer disease duration (SMD_{>18 months from onset} −0.37, 95% CI −0.95 to 0.22; SMD_{≤18 months from onset} −0.24, 95% CI −0.4 to −0.99). There was a high degree of heterogeneity (I ² = 79.3%) in ALS patients exhibiting longer disease duration. In contrast, ALSFRS‐R, site of disease onset, UMN score or gender (proportion of males) were not significant predictors of a relationship between mean ICF and ALS (Table 3).

Motor evoked potential amplitude

The MEP amplitude, expressed as the percentage of the compound muscle action potential response, has been increased in ALS patients signifying cortical hyperexcitability [7]. Increased MEP amplitude was significantly associated with ALS, with evidence of moderate heterogeneity (0.57, 95% CI 0.44–0.71, p < 0.001; Q = 36.25, p = 0.01; I ² = 49%; Figure 5). No significant funnel plot asymmetry was evident for MEP amplitude (Egger's test slope intercept 0.09, p = 0.63).

FIGURE 5.

Forest plot depicting effect sizes for individual studies and an overall effect size for motor evoked potential (MEP) amplitude. The dotted line represents an absence of significant difference in MEP amplitude between amyotrophic lateral sclerosis patients and controls.

Meta‐regression analysis indicated that disease duration, ALSFRS‐R, site of disease onset, UMN score or gender (proportion of males) were not significant predictors of the relationship between MEP amplitude and ALS (Table 3).

Resting motor threshold

Resting motor threshold (RMT) was assessed using the threshold tracking technique across all studies [14]. The SMD was −0.08 (95% CI −0.26 to 0.11, p = 0.40; Q = 75.59, p < 0.001; I ² = 77%; Figure 6), suggesting that RMT poorly discriminated between ALS and controls. Significant effect size heterogeneity was evident, with effect size being markedly different in one study on Japanese ALS patients [34]. Whilst exclusion of this study increased the SMD (−0.14, 95% CI −0.28 to 0.10, p = 0.07; I ² = 62%), the SMD remained non‐significant.

FIGURE 6.

Forest plot depicting effect sizes for individual studies and an overall effect size for resting motor threshold (RMT). The dotted line represents an absence of significant difference in RMT between amyotrophic lateral sclerosis patients and controls.

Meta‐regression analysis suggested that disease‐onset site was a predictor of RMT change in ALS (Table 3). Subgroup analysis disclosed that the effect size was significantly greater in studies reporting a lower (SMD_<30% 0.64, 95% CI 0.4–0.88) compared to those reporting a higher (SMD_≥30% 0.55, 95% CI −0.37 to 0.73; I ² = 57.6%) proportion of bulbar‐onset ALS patients, suggesting that RMT was higher in limb‐onset ALS. In contrast, disease duration, ALSFRS‐R, UMN score or gender (proportion of males) were not significant predictors of the relationship between mean MEP amplitude and ALS (Table 3).

DISCUSSION

The present meta‐analysis, incorporating over 2500 participants, established diagnostic utility of threshold tracking TMS measures of cortical hyperexcitability in differentiating ALS from controls. Reduction of mean SICI (between ISIs 1 and 7 ms) exhibited the highest SMD, suggesting that this TMS measure was the most sensitive TMS biomarker of cortical hyperexcitability in ALS, thereby exhibiting greatest diagnostic utility. The reduction of SICI was independent of disease duration, degree of functional disability, site of disease onset, gender and racial differences. Abnormalities in other TMS measures were in keeping with cortical hyperexcitability, with reduction of CSP duration and increased MEP amplitude exhibiting greater utility than ICF increase or RMT reduction. Incorporation of SICI into commercially available TMS devices could provide an important diagnostic aid, particularly in the early stages of ALS. At a pathophysiological level, the meta‐analysis indicated that motor cortical disinhibition is a more important mechanism for cortical hyperexcitability development, suggesting a proclivity for dysfunction of inhibitory interneuronal circuits and thereby focusing therapeutic strategies towards inhibitory cortical circuits.

Short interval intracortical inhibition (SICI) is a biomarker of cortical inhibitory interneuronal circuits acting via GABA_A receptors and exerting inhibitory effects through synaptic mechanisms, combined with axonal refractoriness and, potentially, shunting inhibition achieved by the opening of channels in proximal dendrites targeted by incoming afferents [18, 56, 57]. Cortical hyperexcitability, heralded by the reduction of SICI, has been identified as an early and intrinsic feature of ALS [2, 3, 4, 58, 59], preceding lower motor neuron dysfunction [46] and associated with an adverse prognosis [42], disease evolution [9, 11, 33] and development of clinical features including the split‐hand phenomenon [10]. Dysfunction or degeneration of the interneuronal GABAergic circuits was postulated to underlie SICI reduction in ALS [6], and thereby development of cortical hyperexcitability. Whilst SICI reduction may represent a compensatory response to lower motor neuron degeneration [2], the findings of normal cortical excitability in neuromuscular mimic disorders of ALS [4, 7, 19], as well as partial and transient normalization of SICI with institution of riluzole therapy (proposed anti‐glutaminergic agent) [43, 49], would argue against such an explanation.

In the current meta‐analysis, a significant reduction of mean SICI (between ISIs 1 and 7 ms) was evident in ALS, with a higher SMD (−0.994, 95% CI −1.12 to −0.873) compared to other TMS parameters of cortical hyperexcitability. Meta‐regression analysis suggested that disease duration, degree of functional disability, site of disease onset, gender or racial background did not significantly influence the relationship between SICI reduction and ALS. Consequently, it could be implied that SICI reduction was an early feature in ALS, consistently evident across different disease stages (as defined by disease duration and level of functional disability), as well as gender, racial background and site of disease onset. Previous studies have reported a greater reduction of SICI in ALS patients with longer disease durations (>17 months) [33, 38], suggesting a progressive dysfunction of cortical GABAergic inhibitory interneuronal circuits. A likely explanation for discordant findings may relate to shorter disease duration (median 15.3 months, range 11–17.5 months) in the current study, whilst not including patients with longer disease durations (>17 months). Underscoring this fact was the finding of a comparable SICI reduction across the King's stages of disease, in which the median disease duration was comparable in patients with disease durations of less than 17 months (stage 1, 12 months; stage 2, 10 months; and stage 3, 12 months) [38].

Reduction in CSP duration was also in keeping with the presence of cortical disinhibition in ALS and in keeping with previous studies [18]. The SMD (−0.655, 95% CI −0.808 to −0.501) was smaller compared to SICI, suggesting that CSP reduction was not as robust a biomarker of cortical hyperexcitability, a finding in keeping with previous studies [7]. The reduction of CSP duration in ALS is mediated by dysfunction of longer latency inhibitory GABA_B circuits as well as concomitant dysfunction spinal inhibitory circuits [60]. Alteration of CSP mediating inhibitory circuits appears to be an early feature of ALS, in keeping with a previous study, which did not evolve with disease progression or functional status [38]. The reason for significant heterogeneity of CSP duration might be multifactorial, relating to muscle fatigue during contraction and inconsistent determination of the onset of CSP. An explanation for greater utility of SICI reduction as a biomarker of cortical hyperexcitability remains to be fully elucidated, although it may reflect greater vulnerability of the shorter latency motor cortical GABAergic circuits in ALS.

An increase in cortical facilitation also contributes to the development of hyperexcitability in ALS, as evident by higher MEP amplitudes and ICF. The MEP amplitude reflects density and excitability of corticomotoneuronal projections onto spinal motor neurons, increased by adrenergic neurotransmission as well as sodium and calcium conductances [56]. ICF may appear to be mediated by distinct cortical facilitatory circuits [61] or through physiological mechanisms operating at a spinal level [62]. The increase in MEP amplitude and ICF did not change with disease duration or functional status, suggesting that these TMS changes become evident during the early stages of ALS. The SMD for MEP amplitude and ICF were smaller compared to SICI and CSP duration, suggesting that cortical inhibition was a greater contributor to development of hyperexcitability than increased facilitation, and in keeping with previous studies undertaken in smaller cohorts [7]. The discordance between measures of cortical inhibition and facilitation could be related to higher MEP amplitude variability [63] and inconsistent identification of ICF [64], suggesting an overall poorer reproducibility of these TMS measures.

Alternatively, the findings achieved through interrogation of cortical motor pathways could be related to a proclivity for dysfunction and degeneration of cortical inhibitory circuits in ALS. Specifically, dysfunction of cortical inhibitory pathways could represent a key pathophysiological mechanism that underlies neurodegeneration in ALS, heralded through the manifestation of cortical hyperexcitability and the ensuing lower motor neuron degeneration. Recent studies have identified the loss of Lamp5 (novel regulator of neuronal hyperexcitation critical for survival of distinct interneuronal populations) in Alzheimer's disease, resulting in dysfunction of cortical networks that are linked to cortical hyperexcitability and neurodegeneration [65]. Assessing for the dysfunction of Lamp5 and related modulators of cortical inhibitory interneuronal function, with a view to unlocking the initial pathogenic event in ALS, may further inform pathophysiological insights into neurological disease and lead to development of novel therapeutic strategies.

Resting motor threshold (RMT) reflects excitability of corticomotoneuronal fibres projecting onto spinal motor neurons that innervate a target muscle [18]. Interestingly, RMT was not significantly reduced in ALS patients in the current meta‐analysis, although subgroup analysis suggested a significant increase in limb‐onset patients. A possible explanation may relate to variability in RMT across different studies, reflecting clinical heterogeneity of ALS. Specifically, whilst some studies reported RMT reduction as an early feature of ALS [8, 21], others have documented normal [2, 3, 9, 21] or increased RMT [22, 23], with the latter findings reflecting UMN degeneration [21]. The effect size in one study was markedly different [34], and omission of this study from the meta‐analysis did not appreciably change the overall effect size (−0.135, 95% CI −0.284 to 0.013, p = 0.074; Q = 49.80, p < 0.001; I ² = 62%). Subgroup analysis disclosed a significant effect size in limb‐onset ALS, suggesting that increased RMT is a more reliable cortical hyperexcitability biomarker in the limb‐onset group of patients. Whilst RMT may be influenced by use of neuromodulating agents, such as Na⁺ channel blockers, this did not account for the findings as none of the studies reported use of neuromodulating agents that could impact RMT values. Rather, the findings could reflect the pathophysiological heterogeneity across the ALS phenotype.

The current meta‐analysis included studies utilizing the threshold tracking TMS technique with serial ascending ordering of ISIs [14]. TMS studies utilizing the constant stimulus method were excluded due to methodological differences, rendering direct comparisons of the paired‐pulse physiological measures difficult. Additionally, the parallel threshold tracking TMS method reported highest diagnostic utility at ISIs of 2.5 ms [58, 59], contrasting with the serial ascending method where mean SICI (between ISI 1 and 7 ms) was reported to exhibit highest diagnostic utility [7]. The differences in area under the curve were subtle, at least as far as the typical ALS phenotype is concerned, and whether these were significantly different could not be determined for the parallel tracking TMS papers as the 95% CIs were not reported. Notwithstanding this, the varied findings could relate to differences in methodology, including the use of different coils (figure of eight vs. circular coil) and recording sites (first dorsal interosseous vs. abductor pollicis brevis muscle), thereby precluding direct comparisons between the parallel tracking method and studies included in the meta‐analysis. Large multicentre studies, comparing the three TMS methodologies, may be required to determine the diagnostic utility of different SICI components in ALS. It should be acknowledged that both constant stimulus [2, 66, 67, 68, 69, 70] and parallel threshold tracking [58, 59] TMS techniques have reported significant SICI reduction in ALS, in keeping with the current meta‐analysis and suggesting that similar cortical disinhibition contributed to ALS pathophysiology. Separately, it is accepted that use of aggregate data rather than individual patient data in the current meta‐analysis could have impacted the findings. Whilst equivalence of summary effects between the aggregate and individual patient data methods has been reported [71], thereby arguing against an impact of the aggregate based methodology on the findings, it should be acknowledged that the trajectory of TMS measures (SICI, ICF, MEP amplitude, CSP duration and RMT) are likely to be variable at an individual patient level, thereby impacting on diagnostic utility. An individual patient data meta‐analysis may help clarify the utility of the various TMS parameters at a patient level.

From a clinical perspective, the present study established that reduction in mean SICI (between ISIs 1 and 7 ms) exhibited the highest SMD, as assessed from a large cohort of ALS patients (>1500) and controls (>1100), implying greater clinical diagnostic utility compared to other TMS biomarkers. Integration of the paired‐pulse TMS paradigm into commercially available devices may facilitate ALS diagnosis by enabling measurements of SICI in a clinical setting, although confirmatory multicentre studies are still required. Additionally, TMS measures of cortical hyperexcitability could also serve as surrogate biomarkers of cognitive and behavioural dysfunction in ALS and ALS–frontotemporal dementia overlap syndrome [35, 36]. Combining SICI with other TMS measures, including CSP duration, MEP amplitude and ICF, could also aid in diagnostic utility, from both motor and cognitive aspects, as well as obviating confounding effects such as neuromodulating medications. If confirmed, clinical application of cortical excitability measurements may lead to earlier commencement of appropriate therapeutic agents in a multidisciplinary clinical environment, as well recruitment into clinical trials [72].

AUTHOR CONTRIBUTIONS

Steve Vucic: Conceptualization; investigation; writing – original draft; visualization; validation; methodology; software; formal analysis; project administration; data curation; supervision; resources. Mana Higashihara: Conceptualization; investigation; writing – original draft; methodology; formal analysis; data curation. Nathan Pavey: Conceptualization; investigation; methodology; validation; writing – review and editing; formal analysis; software; data curation. Parvathi Menon: Conceptualization; writing – review and editing; resources; methodology. Mehdi van den Bos: Conceptualization; writing – review and editing; formal analysis; data curation. Kazumoto Shibuya: Conceptualization; writing – review and editing; formal analysis. Satoshi Kuwabara: Conceptualization; investigation; methodology; writing – review and editing. Matthew Kiernan: Writing – review and editing; conceptualization; supervision. Masayoshi Koinuma: Conceptualization; investigation; writing – review and editing; methodology; formal analysis; data curation; supervision.

CONFLICT OF INTEREST STATEMENT

The authors declare the following funding sources that did not influence the study. PM recipient of the MND Research Australia (Betty Laidlaw Research Award) fellowship; SV recipient of the National Health and Medical Research Council ideas grant (grant no. 2010812); MCK was supported by an NHMRC Practitioner Fellowship (1156093). There are no other conflicts of interest.

Higashihara M, Pavey N, Menon P, et al. Reduction in short interval intracortical inhibition from the early stage reflects the pathophysiology in amyotrophic lateral sclerosis: A meta‐analysis study. Eur J Neurol. 2024;31:e16281. doi: 10.1111/ene.16281

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available from the corresponding author upon reasonable request.