Read at the level of the cohort, the long-term data are genuinely good. In a single-surgeon series of 34 patients with a mean follow-up of 56.9 months, the mean tremor score fell from 3.27 to 0.64 on stimulation, handwriting from 2.94 to 0.89, corresponding to an 80.4% improvement in tremor and 69.7% in handwriting; in the 12 patients compared between roughly 57 and 91 months, scores did not significantly change (Zhang et al., J Neurosurg, 2010). A large single-center experience of thalamic DBS across Parkinson disease, essential tremor, and dystonia similarly documented sustained tremor control at long follow-up (Cury et al., Neurology, 2017). The aggregate message is that Vim-DBS is not a short-lived intervention — most patients retain meaningful benefit for years.
The signal that complicates that reassurance sits in the stimulation-parameter data. The same long-term series that report sustained benefit also report that stimulation settings are progressively escalated, with a significant increase in parameters over the first five years (Zhang et al., J Neurosurg, 2010), and cohorts followed to a mean of 7.7 years describe a loss of efficacy whose mechanism is "not well understood" alongside steadily adjusted programming (Rodríguez Cruz et al., Mov Disord Clin Pract, 2016). Sustained outcome and rising energy demand coexist: the benefit is maintained, but maintaining it takes more.
The response is also not uniform across tremor distribution. Distal limb (hand) tremor is where Vim-DBS responds most completely; midline and axial components — notably head and voice tremor — often respond less completely. When a patient reports that their handwriting is still helped while their head or voice tremor remains prominent, that pattern can be misread as global habituation when it may instead reflect a distribution-specific limitation of the therapy rather than a benefit that has waned over time.
The waning of an initially excellent DBS response over a variable interval is the clinical entity at issue, and the terminology has shifted. Historically it was called "tolerance," borrowing a pharmacological word; the movement-disorders literature now often prefers "habituation" precisely to avoid implying a receptor-level mechanism that has not been established (Fasano & Helmich, Mov Disord, 2019). Whichever label is used, the observation is the same: a patient who did well initially returns with tremor that stimulation no longer controls as it once did, and the deterioration is more than can be explained by the natural course alone.
Two features make it clinically distinctive. First, it can appear early — habituation frequently presents as a substantial loss of the initial benefit within a few months of stimulation onset, not only after many years (Peters & Tisch, Front Neurol, 2021). Second, it is partially reversible in a way disease progression is not: a proportion of the lost benefit can be recovered by reprogramming, which is itself a diagnostic clue as much as a treatment.
The honest answer is that the reported prevalence is almost uninterpretable as a single figure. Across studies, waning benefit has been reported in anywhere from 0% to 73% of patients (Fasano & Helmich, Mov Disord, 2019). That 73-point spread is not a disagreement about biology; it is an artifact of definition. Studies differ on what counts as habituation (any rise in tremor score? a threshold drop from peak benefit? a need to reprogram?), over what interval, and measured with which scale — and habituation quantification has accordingly produced conflicting reports (Peters & Tisch, Front Neurol, 2021).
The most granular attempt to pin it down is a repeated-measures meta-analysis of 533 patients from 18 studies, evaluating Vim-DBS efficacy at intervals out to four or more years. Its findings are usefully unflattering to a simple narrative: the motor subscore showed no significant loss of efficacy, while hand-function and activities-of-daily-living scores did decline — and the analysis attributed the hand-function loss to DBS tolerance and the daily-living loss to both tolerance and disease progression (Bai et al., CNS Neurosci Ther, 2022). In other words, whether Vim-DBS "loses efficacy" depends heavily on which outcome you measure, which is exactly why the prevalence figure is so elastic.
The strongest single piece of evidence that the decline is not merely progression or placebo comes from an observer-blinded long-term study. Vim-DBS improved the Tremor Rating Scale in both the short and long term, but the magnitude of the stimulation effect was negatively correlated with time since surgery (ρ = −0.78, p < 0.001), and long-term worsening was more profound in the stimulation-ON than in the stimulation-OFF condition — a pattern the authors read as habituation to stimulation rather than disease progression alone. Only about a third of patients showed a rebound effect 60 minutes after the stimulator was switched off, and the study concluded plainly that Vim-DBS loses efficacy over the long term (Paschen et al., Neurology, 2019). That ON-versus-OFF dissociation is the methodological crux: if worsening were purely progression, it should be similar with stimulation on or off.
Apparent loss of DBS benefit is best treated as a differential, not a diagnosis. The contributors that the literature identifies include, in rough clinical order:
Placebo effects, loss of the micro-lesional effect, disease-related progression, suboptimal stimulation, and stimulation-related side effects may all contribute to the loss of sustained therapeutic effect, and they are not mutually exclusive in a given patient (Peters & Tisch, Front Neurol, 2021). This is why "the DBS stopped working" is a starting point for evaluation, not a conclusion — and why the meta-analytic finding that motor scores can hold while functional scores fall matters at the bedside (Bai et al., CNS Neurosci Ther, 2022).
Because a share of apparent habituation is programming-responsive, the first response to waning benefit is systematic reprogramming rather than resignation: adjusting active contacts, voltage, frequency, and pulse width, and in some centers trialing stimulation holidays, are the described levers (Fasano & Helmich, Mov Disord, 2019). Efforts to predict who will habituate are advancing — objective neurophysiological measures have been proposed as predictors of "early tolerance" to Vim-DBS, which could eventually inform programming or target decisions before benefit is lost (Merchant et al., Clin Neurophysiol, 2018).
Target selection is the other lever, and it is where the patient's instinct to ask "VIM or PSA?" is clinically apt. The ventral intermediate nucleus is the standard target, but the caudal zona incerta (part of the posterior subthalamic area) has favorable long-term data as an alternative, with sustained tremor reduction on the Essential Tremor Rating Scale at three-to-five-year follow-up in a dedicated cohort (Fytagoridis et al., J Neurol Neurosurg Psychiatry, 2012). Whether one target habituates less than another is not settled, but target choice demonstrably shapes both efficacy and the re-operation options available later.
The long-term data reshape how the intervention should be framed before surgery. The accurate message to a candidate is not "this will fix your tremor permanently" but "this is a highly effective, durable therapy that most patients keep for years — and one that will need periodic adjustment, may demand rising stimulation over time, and can partially wane in a minority." That framing is supported rather than undercut by the evidence: aggregate benefit is large and sustained (Zhang et al., J Neurosurg, 2010; Cury et al., Neurology, 2017), yet parameter escalation and a measurable time-dependent decline are the norm, not the exception (Paschen et al., Neurology, 2019).
Practically, that argues for structured long-term follow-up rather than discharge after programming stabilizes. Because functional and daily-living scores can erode while a bedside motor exam still looks controlled (Bai et al., CNS Neurosci Ther, 2022), follow-up that tracks patient-reported function — not only clinician tremor ratings — is more likely to catch early waning when reprogramming can still recover it. And when a patient does return with lost benefit, the ON-versus-OFF comparison, an imaging review of lead position, and a systematic reprogramming trial should precede any conclusion that the tremor has simply progressed beyond the device's reach.
Several questions remain genuinely open and should be represented as such. There is no agreed operational definition of habituation, which is the single largest reason prevalence estimates span 0–73% and studies cannot be pooled cleanly (Fasano & Helmich, Mov Disord, 2019; Peters & Tisch, Front Neurol, 2021). Whether habituation reflects a true central-nervous-system adaptation, an artifact of progression plus fixed programming, or a mixture that differs by patient is unresolved, and the meta-analytic dissociation between preserved motor scores and declining functional scores suggests the answer is not uniform (Bai et al., CNS Neurosci Ther, 2022). And there is no validated way, at the individual level, to prospectively identify the patient who will habituate, though neurophysiological predictors are an active area (Merchant et al., Clin Neurophysiol, 2018). The defensible clinical stance is neither that DBS reliably fails nor that habituation is a myth, but that long-term benefit is real, common, and maintainable — while requiring active management and honest pre-operative counseling that the device is a durable therapy, not a fixed one.