Cerebellar lesions
The cerebellar SVV is more about variability than magnitude — and that is its diagnostic fingerprint.
Checking for audio...
The cerebellar SVV is the one where the mean lies to you. Mean values are often within normal limits. The variability — the standard deviation across trials — is the diagnosis.
The cerebellum, particularly the nodulus, uvula, and flocculus, processes graviceptive information and calibrates the central estimate of vertical. Cerebellar damage degrades the precision of that estimate without consistently biasing its direction.
On bucket test, a cerebellar patient's ten trials might read minus six, plus three, minus two, plus five, minus four, and so on. The mean is zero. The standard deviation is four degrees. The patient looks normal until you compute the spread.
Report both mean and SD on every SVV report. Cerebellar SVV is missed when you only report the average.
Localising patterns. Nodulus and uvula lesions give the largest variability, sometimes with associated periodic alternating nystagmus. Cerebellar nuclei lesions show moderate variability with a small contraversive mean. Flocculus lesions produce variability without consistent tilt, alongside gaze-evoked nystagmus and saccadic pursuit.
Cerebellar SVV does not localise reliably to a side. It does not give a complete OTR. It does not produce a measurable head tilt. It is a variability sign, not a direction sign.
Expected SVV signature
| Feature | Finding |
|---|---|
| Mean tilt | Often within ±3°; can be larger with focal nodulus/uvula lesions |
| Direction | Variable, often contraversive in cerebellar nuclei lesions |
| Trial-to-trial SD or IQR | Increased — the defining feature |
| OTR triad | Usually incomplete; head tilt and skew often absent |
The cerebellum (especially the nodulus, uvula, and flocculus) processes graviceptive information and calibrates the central estimate of vertical. Cerebellar damage degrades the precision of that estimate more than its accuracy, so the patient’s settings scatter widely around the true vertical rather than tilting consistently to one side[2].
Magnitude vs variability — report both
Reporting only the mean SVV will miss cerebellar lesions. A mean of −0.5° looks normal until you notice the standard deviation across 10 trials is 4°. The patient placed the line at −6°, +3°, −2°, +5°, −4°… all consistent with the same diagnosis: defective central calibration of vertical.
Whenever you report SVV, report the SD or IQR alongside the mean. For focal cerebellar lesions (PICA infarct sparing the medulla, multiple sclerosis plaque in the cerebellar peduncle, paraneoplastic cerebellar degeneration), the variability is often the only quantitative SVV finding.
Localising patterns
- Nodulus / uvula — large variability, can be associated with periodic alternating nystagmus. Often paraneoplastic or post-infectious.
- Cerebellar nuclei (dentate, fastigial) — moderate variability with small contraversive mean tilt.
- Flocculus — variability without consistent tilt; gaze-evoked nystagmus and impaired smooth pursuit dominate the clinical picture.
Companion findings
- Gaze-evoked nystagmus, often horizontal and direction-changing
- Impaired smooth pursuit; saccadic pursuit
- Truncal ataxia, gait ataxia, dysdiadochokinesia
- Normal head impulse test (HIT) — cerebellar lesions do not impair the VOR gain
- Down-beating nystagmus with floccular involvement