Clinicians treating hereditary ataxia may one day get faster answers from MRI scans after Monash University researchers demonstrated that geometric patterns on the brain's surface can flag Friedreich ataxia and related disorders.
The team merged structural imaging — the hospital scans neurologists already order — with functional MRI maps of activity, then analysed both through mathematical shapes called geometric eigenmodes. Applied to more than 500 participants across three cohorts, the model spotted differences between healthy volunteers and people with Friedreich ataxia, spinocerebellar ataxia types 1 and 3, using 200 shape-based features spanning the cortex and cerebellum.
Why diagnosis drags today
Hereditary ataxias are progressive conditions affecting coordination, speech and gait. Patients often endure years of tests because symptoms overlap with other neurological diseases and because individual hospitals lack rare-disease expertise. Genetic confirmation remains the gold standard, but MRI clues could prioritise who needs urgent sequencing and monitor progression between visits.
Deep-learning classifiers in the study separated Friedreich ataxia from controls and distinguished some subtypes. Researchers caution the pipeline is experimental: it needs larger, multi-site trials before regulators accept it as a diagnostic aid rather than a research tool.
How the method works
Geometric eigenmodes treat the brain's folded surface like a vibrational structure, highlighting where atrophy or functional disconnects concentrate. Combining anatomical thinning with activity deficits produced sharper boundaries than either scan type alone. That matters for Australia because MRIs are reimbursed through Medicare yet neurologists still eyeball films subjectively.
If validated, hospitals could run the algorithm on existing archives, building population baselines without new hardware. Regional centres where patients travel hours to see an ataxia specialist would benefit most, cutting repeat imaging and misdiagnosis costs.
Path to the clinic
Monash collaborators include international partners who supplied independent cohorts, a key step to prove the signatures generalise beyond one machine or demographic. Next stages involve prospective trials tracking patients annually to see whether geometric scores predict falls, wheelchair reliance or treatment response.
For families waiting on a label, the science does not replace genetic counselling. It offers a potential triage layer — much like AI tools emerging in oncology pathology — that respects existing clinical workflows. Australian radiology groups will watch reimbursement discussions closely; a certified algorithm could attract item numbers if outcomes data firm up.
Australian context
Ataxia clinics in Melbourne and Perth already participate in natural-history registries that could feed future training data. Monash-affiliated hospitals in Clayton and Moorabbin see steady referral volumes from regional GPs uncertain about hereditary versus acquired causes. Reducing repeat imaging would free capacity in a health system where MRI slots are rationed by triage committees.
Neurology training programs may incorporate the geometric approach into radiology fellowships if peer-reviewed publication follows. Until then, families should expect standard care pathways while researchers pursue grant funding for multi-year follow-up.