Published by Unseen Progress, an independent publisher of caregiver research. Last reviewed 2026-05-10. Part of the sensory processing research overview.
Short answer. A sensory diet is a personalised, scheduled set of sensory activities prescribed by an occupational therapist to support a child's regulation and tolerance development. The term was coined by Patricia Wilbarger in the 1980s. A well-designed sensory diet has four properties: it is per-domain (auditory, tactile, vestibular, proprioceptive, oral, etc.), scheduled (predictable times, not improvised), graduated (intensity changes as the nervous system adapts), and measured (parents track adherence and outcomes for at least 90 days before evaluating). Most failed "sensory diets" violate one or more of these properties — usually the schedule and the measurement.
A sensory diet is not a list of fun sensory activities the child enjoys. It is a clinical prescription, structured the way a medication regimen is structured, with a stated mechanism for each activity, a frequency, and a duration. Schaaf and Mailloux (2015) emphasise that the design is what carries the evidence base, not the activities in isolation — swinging is therapeutic when it is part of a targeted vestibular protocol, and arbitrary when it isn't.
The four properties below are what separates the protocols that produce 3–12 month tolerance gains (Schaaf & Mailloux, 2015) from the ones that fizzle out by week six.
Miller et al. (2007) distinguish modulation difficulties across at least eight sensory domains: auditory, tactile, visual, vestibular, proprioceptive, oral, olfactory, and interoceptive. Most sensory-sensitive children have problems in two to four of these, not all eight, and the affected domains are different from child to child. A diet that doesn't name which domains it is targeting is a diet you can't evaluate.
Concretely, a per-domain diet looks like:
A diet that simply lists "swing, brush, weighted vest, chew necklace" without per-domain mechanism is what most parents are handed and what most fail to sustain — because there's nothing to measure against.
Wilbarger's original framing of a sensory diet (1991) is structurally analogous to a meal plan: predictable times, predictable doses, integrated into the day. Improvised sensory input — "she seems dysregulated, let's try the swing" — is reactive co-regulation, which is helpful in the moment but is not what produces multi-month tolerance gains.
The research-backed structure is to schedule sensory-diet activities at known regulation pressure points — morning wake-up, before school, after school, before transitions, before bed — and at intervals that match the child's regulation half-life as observed by the OT. The schedule is what allows the protocol to be evaluated. Without it, the parent cannot tell whether the diet is working or whether the child happened to have a calmer week.
Sensory tolerance expands by graduated exposure within a calm context (Schaaf & Mailloux, 2015). A static diet — same brushing pressure, same swing duration, same food textures for nine months — produces an initial regulation benefit and then plateaus. The OT's role is to step the intensity up as the nervous system adapts: longer tolerance windows, more intense vestibular input, novel fabrics, new food textures introduced one at a time.
The most common parent error in this property is not failure to escalate — most parents under-escalate, which is the safer error — but escalating without OT input, which can re-trigger a child whose nervous system was almost ready but not quite. Escalation decisions are clinician decisions; the parent's job is to bring data that lets the clinician decide.
This is the property most diets fail on, and the one the research speaks to most directly. Schaaf and Mailloux (2015) report meaningful tolerance changes on a 3–12 month timescale. The first 4–8 weeks typically show no visible change even when the underlying nervous system is beginning to retune. A parent who is evaluating from memory at week six will conclude the diet isn't working at exactly the point it usually is.
Concrete measurement looks like:
Parents who run a 90-day measurement window are the ones who get the data the OT needs to step the protocol forward. Parents who run on memory alone are the ones who, in good faith, abandon protocols that were starting to work.
Most paediatric OT practices follow some variant of this sequence:
1. Sensory profile assessment. Dunn's Sensory Profile (Dunn, 1999, 2014) or the Sensory Processing Measure identifies which domains and which subtypes (over-responsive, under-responsive, seeking) are present. 2. Hypothesis-driven activity selection. For each affected domain, the OT picks activities with a stated mechanism — proprioceptive heavy work for vestibular regulation, deep pressure for tactile defensiveness, joint compressions for body-awareness deficits. 3. Schedule integration. Activities are placed at the pressure points of the child's day, in collaboration with the parent so the schedule is realistic for the household. 4. Adherence tracking and review cycle. The OT and parent agree on what data the parent will bring back, and on a review cadence — typically every 2–4 weeks initially, every 6–8 weeks once the diet is stable.
Diets built this way are not glamorous. They are short, specific, measurable, and boring on purpose — because the boring ones are the ones parents can actually sustain through the months it takes to see results.
---
Unseen Progress publishes long-form caregiver research and builds research-backed daily trackers for the families covered. See the full sensory processing research overview for the complete framework.