Introduction

Chronic ataxias represent a diverse group of rare, progressive neurological disorders marked by significant impairment in balance and coordination, most often due to cerebellar dysfunction or disruptions across broader neural networks. These conditions are frequently rooted in complex genetic mutations and are notoriously difficult to diagnose. Symptomatically, they overlap with other movement disorders, and their underlying genetic architecture—spanning more than 300 causative genes—adds further complexity, with broad phenotypic variability confounding clinical assessments and delaying accurate diagnoses.

Traditional diagnostic pathways—including clinical evaluations, MRI, and targeted genetic testing—often fall short. Many patients remain undiagnosed or misdiagnosed for years, missing critical opportunities for timely intervention, appropriate management, and family counseling. Against this backdrop, artificial intelligence (AI) and machine learning (ML) are emerging as powerful tools in neurology. These technologies offer scalable, reproducible analysis of multimodal data, enabling the detection of subtle disease patterns that elude human observation. In chronic ataxias, AI holds the potential to close diagnostic gaps, accelerate disease recognition, and guide individualized care strategies—raising a pivotal question: can AI set a new standard for precision diagnosis in these challenging neurogenetic disorders?

What Are Chronic Ataxias?

Chronic ataxias are neurological disorders characterized by persistent motor coordination deficits arising from cerebellar or related neural dysfunction. Unlike acute ataxias—which typically result from stroke, infection, or trauma—chronic ataxias progress insidiously over time. Their diagnosis is complicated by gradual symptom onset, clinical overlap with other neurological conditions, and wide etiological variability, ranging from inherited to acquired causes.

Major forms include spinocerebellar ataxias (SCAs), Friedreich’s ataxia, multiple system atrophy (MSA), and ataxia-telangiectasia, each associated with significant genetic heterogeneity. Over 90 genes have been linked to autosomal recessive cerebellar ataxias alone, contributing to a diagnostic landscape that is both fragmented and time-consuming. It is not uncommon for patients—particularly those with rare subtypes—to wait over five years from symptom onset to confirmed diagnosis.

Diagnostic strategies typically involve neurological examination, detailed family history, neuroimaging, and stepwise genetic testing through gene panels or next-generation sequencing. However, these approaches are limited by the subjective nature of clinical assessments, inconsistent access to comprehensive genetic tools, and data fragmentation across specialties. Even after genetic testing, studies show that up to 71% of patients may remain undiagnosed, underscoring the urgent need for broader genetic screening and more integrative diagnostic platforms.

AI in Action: A New Frontier for Diagnosing Ataxia

Artificial intelligence is reshaping the diagnostic landscape of chronic ataxias by applying machine learning and deep learning techniques across multiple data domains—including neuroimaging, electronic medical records (EMRs), and genomics. Convolutional neural networks (CNNs), for example, have demonstrated impressive accuracy in identifying disease-specific MRI patterns and differentiating among various SCA subtypes based on neuroanatomical signatures.

Natural language processing (NLP) algorithms further expand diagnostic capabilities by extracting complex phenotypic patterns and clinical insights from unstructured EMR data. Meanwhile, decision-tree-based classifiers that integrate clinical presentations with genomic data offer advanced diagnostic stratification tools. Recent breakthroughs underscore AI’s potential: one study reported a virtual diagnostic assistant, combining algorithms and decision trees, achieving a diagnostic accuracy of 90.9%—outperforming both neurologists and GPT-4 in identifying a wide spectrum of chronic ataxias, including early-stage Friedreich’s ataxia.

Other AI applications have achieved 86–88% accuracy in distinguishing patients from healthy controls using inertial measurement unit data. Machine learning models have also shown significant correlation with clinical scales in predicting disease progression and severity, offering new avenues for patient monitoring.

Reference:

  1. Rudaks LI, Yeow D, Ng K, Deveson IW, Kennerson ML, Kumar KR. An Update on the Adult-Onset Hereditary Cerebellar Ataxias: Novel Genetic Causes and New Diagnostic Approaches. Cerebellum. 2024 Oct;23(5):2152-2168. doi: 10.1007/s12311-024-01703-z. Epub 2024 May 18. PMID: 38760634; PMCID: PMC11489183.

  2. AbuAlrob MA, Mesraoua B. Harnessing artificial intelligence for the diagnosis and treatment of neurological emergencies: a comprehensive review of recent advances and future directions. Front Neurol. 2024 Oct 11;15:1485799. doi: 10.3389/fneur.2024.1485799. PMID: 39463792; PMCID: PMC11502371.

  3. Lacorte E, Piscopo P, Sagliocca L, Vignatelli L, Taruscio D, Vanacore N and The Ataxia Guideline Development Working Group (GDWG) (2022) The National Italian Guidelines on the diagnosis and treatment of children with pediatric ataxias. Front. Neurol. 13:971781. doi: 10.3389/fneur.2022.971781

  4. Lopergolo D, Rosini F, Pretegiani E, Bargagli A, Serchi V and Rufa A (2024) Autosomal recessive cerebellar ataxias: a diagnostic classification approach according to ocular features. Front. Integr. Neurosci. 17:1275794. doi: 10.3389/fnint.2023.1275794

  5. Shao L, Wang H, Xu J, Qi M, Yu Z and Zhang J (2023) Ataxia-telangiectasia in China: a case report of a novel ATM variant and literature review. Front. Neurol. 14:1228810. doi: 10.3389/fneur.2023.1228810

  6. Jayadev, S., Bird, T. Hereditary ataxias: overview. Genet Med 15, 673–683 (2013). https://doi.org/10.1038/gim.2013.28

  7. Alessandro L, Bianciotti N, Salama L, Volmaro S, Navarrine V, Ameghino L, Arena J, Bestoso S, Bruno V, Castillo Torres S, Chamorro M, Couto B, De La Riestra T, Echeverria F, Genco J, Gonzalez Del Boca F, Guarnaschelli M, Giugni JC, Laffue A, Martinez Villota V, Medina Escobar A, Paez Maggio M, Rauek S, Rodriguez Quiroga S, Tela M, Villa C, Sanguinetti O, Kauffman M, Fernandez Slezak D, Farez MF, Rossi M. Artificial Intelligence-Based Virtual Assistant for the Diagnostic Approach of Chronic Ataxias. Mov Disord. 2025 May;40(5):906-916. doi: 10.1002/mds.30168. Epub 2025 Mar 22. PMID: 40119570.

  8. Vizcarra JA, Yarlagadda S, Xie K, Ellis CA, Spindler M, Hammer LH. Artificial Intelligence in the Diagnosis and Quantitative Phenotyping of Hyperkinetic Movement Disorders: A Systematic Review. J Clin Med. 2024 Nov 21;13(23):7009. doi: 10.3390/jcm13237009. PMID: 39685480; PMCID: PMC11642074.

  9. Fyfe I. AI aces diagnosis of chronic ataxias. Nat Rev Neurol. 2025 Jun;21(6):294. doi: 10.1038/s41582-025-01102-w. PMID: 40374958.