Machine Learning in Amyotrophic Lateral Sclerosis: Achievements, Pitfalls, and Future Directions

Area of Science:

• Neuroscience
• Biomedical Engineering
• Data Science

Background:

• Amyotrophic Lateral Sclerosis (ALS) is a progressive neurodegenerative disease with limited treatments and unclear etiology.
• Accurate diagnostic, monitoring, and prognostic markers are lacking, hindering drug development and clinical trials.
• Disease heterogeneity and patient cohort admixing present significant challenges in ALS research.

Purpose of the Study:

• To systematically review Machine Learning (ML) applications in Amyotrophic Lateral Sclerosis (ALS).
• To provide a clinical-mathematical perspective on ML advancements and future directions in ALS research.
• To critically discuss the limitations of current ML models and offer recommendations for future study designs.

Main Methods:

• Comprehensive literature review of ML initiatives in ALS.
• Analysis of ML model applications in diagnosis, prognosis, and patient stratification.
• Critical evaluation of methodological strengths and weaknesses of existing studies.

Main Results:

• ML techniques have shown success in developing promising diagnostic models for ALS, despite sample size limitations.
• Prognostic models utilizing clinical, biological, and neuroimaging data have been developed for ALS patient stratification.
• Current ML studies in ALS often violate statistical assumptions and lack model justification or stated constraints.

Conclusions:

• Limited sample sizes remain a primary barrier for developing validated ALS biomarkers using ML.
• Integrating multiple clinical, biofluid, and imaging biomarkers can enhance ML model accuracy.
• Optimized ML models and biomarker combinations are crucial for improving ALS clinical trial designs.