Pattern analysis of peripheral-vestibular deficits with machine learning using hierarchical clustering.

Area of Science:

• Neuroscience
• Otolaryngology
• Medical Diagnostics

Background:

• Vestibular organ disorders can cause unique peripheral vestibular loss patterns, aiding diagnosis.
• Neuropathological testing allows classification of vestibular sensor deficits.
• Understanding these patterns is crucial for accurate diagnosis and management of vestibular dysfunction.

Purpose of the Study:

• To review and summarize advances in pattern recognition for unilateral and bilateral vestibular disease.
• To evaluate the utility of hierarchical cluster analysis (HCA) in identifying vestibular loss patterns.
• To explore the association between specific peripheral vestibular loss patterns and underlying disorders.

Main Methods:

• Topical review of published studies by the authors.
• Application of hierarchical cluster analysis (HCA) for pattern recognition in vestibular disease.
• Analysis of vestibular end-organ damage patterns in various patient cohorts.

Main Results:

• Disease-specific patterns of peripheral vestibular loss were identified.
• Anterior-canal sparing was a common feature in aminoglycoside-related bilateral vestibulopathy, Meniere's disease, and vestibular Schwannoma.
• Acute superior-branch vestibular neuropathy showed a pattern involving the horizontal and anterior canals and the utricle.

Conclusions:

• Hierarchical cluster analysis (HCA) effectively facilitates pattern identification in unilateral and bilateral vestibulopathies.
• Vestibular end-organ damage varies extensively across different disorders and patient subgroups.
• Careful selection of clustering algorithms is necessary due to the diversity of available methods.