Electroencephalography (EEG) measures of neural activity and connectivity can be used to reproducibly group patients with amyotrophic lateral sclerosis (ALS) into subphenotypes with distinct clinical patterns and neurophysiologic signatures, according to a study published in Brain.
ALS is a neuromuscular disorder that is mainly characterized by motor system degeneration and associated with cognitive and behavioral change in half of cases. The disorder is known to be heterogeneous with different patterns of disease course and prognosis. However, it’s difficult for clinicians to predict these patterns in advance. Recent research has shown resting-state EEG has the ability to identify different patterns of brain network disruption which are indicative of the underlying disease process.
The objective of the current study was to determine whether EEG can help identify ALS disease subphenotypes and whether different patterns of disruption are predictive of disease outcome.
Researchers conducted a clustering analysis to identify ALS disease subphenotypes and to determine whether different patterns of disruption are predictive of outcome with use of similarity network fusion and spectral clustering.
The patients were enrolled from the National ALS clinic in Beaumont Hospital Dublin within 18 months of their diagnosis. Healthy control individuals included neurologically normal, age-matched participants from a population-based control bank.
The researchers collected EEG data with 128 channels. Initial recording sessions were performed in 95 patients with ALS and 77 healthy controls, and 36 ALS patients had 1 follow-up EEG session after 4 to 6 months.
Patients with ALS had a mean (SD) age of 59.2 ± 11.6 years, a mean disease duration of 21.9 ± 17.5 months, and 69 were male. The patients and controls were matched for age (Mann-Whitney U test, P =.73), but not for sex (Fisher exact test, P <.001).
The researchers identified 4 distinct clusters after analysis of spectral EEG patterns of neural activity and connectivity. The solution of the 4 clusters had high statistical power (0.85 and 0.52, respectively) and a large to medium (0.92 and 0.69, Cliff’s d, respectively) effect size, which suggested reproducible findings, according to the researchers.
Cluster 1 revealed a characteristic increase in β-band spectral power in the frontotemporal network, and clusters 3 and 4 showed decreased power in the same network. Cluster 2 demonstrated a characteristic increase in α-band synchrony in the somatomotor network, cluster 3 a decrease in γ1-band synchrony in the frontotemporal network, and cluster 4 an increase in γ1-band co-modulation in the frontoparietal network.
The EEG abnormalities associated with all 4 clusters were increased comodulation (δ- to α-band oscillations) and decreased synchrony (δ- to β-band) in the somatomotor and frontotemporal brain regions.
Additional analysis showed that each cluster had high accuracy and robustness and remained stable during re-assessment.
The researchers noted that their findings are limited by the single-site design and that alternative solutions with more than 4 clusters are likely to exist, especially if more sophisticated neurophysiological measures were to be included in the clustering analysis.
“We have shown that analysis of network disturbance using multi-dimensional quantitative EEG can identify subgroups within ALS that are not discoverable using standard clinical assessment tools,” they stated. “Each of the subgroups, identified by data-driven clustering, demonstrates a distinct neurophysiological profile that in turn recapitulates a different combination of clinical attributes. These neurophysiological profiles are stable at re-assessment and are associated with different prognostic outcomes.”
Dukic S, McMackin R, Costello E, et al. Resting-state EEG reveals four subphenotypes of amyotrophic lateral sclerosis. Brain. Published online November 17, 2021. doi: 10.1093/brain/awab322