At a glance

Evidence table

Everything on this site in one view. Read it by row: a therapy is rarely at the same level for every condition, and the gap between the best row and the worst is usually where families get into difficulty.

Established care

In clinical trials

Early research only

Not supported by evidence

Layer one

Established care

Early detection and early intervention

Condition	Evidence	Note
Cerebral palsy — detection before 5 months	Established care	International clinical practice guidance supports combining the General Movements Assessment, a structured neurological examination and MRI to identify or strongly suspect cerebral palsy in early infancy.
Cerebral palsy — intervention after early detection	Established care	Guideline-backed. The recommended approach is infant-led, task-specific practice with the family, started as soon as risk is identified rather than waiting for a confirmed diagnosis.
Preterm and high-risk infants generally	Established care	Structured developmental follow-up is standard care in most health systems, and is how most early diagnoses are made.

Goal-directed, task-specific training

Condition	Evidence	Note
Cerebral palsy — motor function	Established care	The most consistently supported approach in cerebral palsy guidance across age groups.
Acquired brain injury	Established care	Same principle, same support.
Genetic and neuromuscular conditions	In clinical trials	Widely applied and biologically sensible; formal trial evidence is thinner in rarer conditions.

Early behavioural and developmental intervention

Condition	Evidence	Note
Naturalistic developmental behavioural interventions (ESDM, PRT, JASPER)	Established care	The best-supported group. These blend behavioural learning principles with developmental science: adult follows the child's lead in play, embeds learning in natural routines, and builds on what the child already initiates.
Parent- and caregiver-mediated programmes	Established care	A therapist coaches the family rather than treating the child directly. Consistently supported, far cheaper, and it puts the hours where the child actually lives.
Behaviour-analytic methods for specific skills	Established care	For toileting, feeding, sleep routines and genuinely dangerous behaviour, structured behavioural methods have good support and a clear, bounded goal.
Traditional intensive EIBI (25–40 hours a week)	In clinical trials	The intensity figure comes from early work and has never been settled by trial. A 2024 analysis of 144 studies found no clear dose–outcome relationship; a reanalysis including IQ reported the opposite, and the two groups still disagree.
DIR / Floortime	In clinical trials	Developmental and relationship-based: the adult joins the child's own play to open and close “circles of communication”. Small randomised trials report gains in caregiver–child interaction; most used unblinded caregiver-reported outcomes, so the evidence is thinner than for NDBI.
Programmes aimed at making a child appear non-autistic	Not supported by evidence	Reducing visible autistic behaviours is not the same as improving a child's life, and is not an outcome worth pursuing on its own.
Aversive procedures	Not supported by evidence	Punishment-based techniques are not acceptable practice and are not part of any contemporary programme worth attending.

Constraint-induced therapy and bimanual training

Condition	Evidence	Note
Unilateral cerebral palsy — hand function	Established care	Strong randomised evidence for both approaches. Guideline-recommended.
Bilateral cerebral palsy	In clinical trials	Bimanual approaches are used; constraint has less rationale when both hands are affected.
Acquired hemiplegia	Established care	Same principle, applied after stroke or injury.

Physiotherapy and strength training

Condition	Evidence	Note
Cerebral palsy — gross motor function	Established care	Function-focused physiotherapy is standard, guideline-supported care.
Strength and fitness	Established care	Progressive resistance training improves strength. Translation into walking function is less consistent.
Passive stretching to prevent contracture	Early research only	Long-held belief, weak evidence. Stretching alone has not been shown to prevent contracture.

Occupational therapy

Condition	Evidence	Note
Cerebral palsy and motor disability	Established care	Goal-directed occupational therapy is guideline-supported.
Autism — daily function and participation	Established care	Supported where goals are functional. Sensory-integration approaches as a treatment for core autism features are a separate and weaker claim.
Developmental coordination difficulties	Established care	Task-oriented approaches have good support.

Speech, language and communication

Condition	Evidence	Note
Augmentative and alternative communication (AAC)	Established care	Well supported. Evidence indicates AAC does not inhibit speech development and is often associated with gains in speech.
Speech and language therapy in cerebral palsy	Established care	Standard care, including for eating and drinking safety.
Language intervention in autism	Established care	Naturalistic developmental behavioural approaches have the strongest support.
Oral-motor exercises to improve speech	Early research only	Non-speech oral exercises have not been shown to improve speech sound production.

Spasticity and tone management

Condition	Evidence	Note
Focal spasticity — botulinum toxin	Established care	Licensed and guideline-supported for focal spasticity in children, as part of an integrated programme with therapy.
Generalised spasticity — intrathecal baclofen	Established care	Established for selected children with severe generalised spasticity, in specialist centres.
Selective dorsal rhizotomy	Established care	Established for carefully selected children, most often ambulant children with spastic diplegia. Selection is the whole question.
Oral antispasticity medication	In clinical trials	Widely used; evidence for functional benefit in children is modest and side effects limit dose.
Serial casting and orthoses	Established care	Supported for specific goals such as improving ankle range or foot position in walking.

Technology-assisted rehabilitation

Condition	Evidence	Note
Robotic gait training / treadmill with body-weight support	In clinical trials	Can deliver high repetition. Evidence of advantage over equally intensive conventional training is not established.
Functional electrical stimulation	In clinical trials	Used for foot drop and to assist grasp; benefit reported, best evidence when combined with active practice.
Virtual reality and active video games	In clinical trials	Improves engagement and can increase practice volume. Effects on function are modest.
Exoskeletons for home use	Early research only	Marketed ahead of the evidence in children.

Named physiotherapy approaches

Condition	Evidence	Note
Bobath / neurodevelopmental treatment (NDT)	In clinical trials	The most widely taught approach in the world. Reviews have not shown it superior to goal-directed, activity-based therapy, and several guidelines have moved away from recommending it as the default.
MEDEK / Cuevas MEDEK Exercises	Early research only	Provocative, gravity-based handling to elicit postural responses. Enthusiastic parent reports, very little controlled evidence.
MAES therapy	Early research only	A newer approach emphasising the child's own problem-solving. Coherent reasoning; published controlled trials are essentially absent.
Anat Baniel Method / NeuroMovement	Early research only	Derived from Feldenkrais, built on gentle movement and attention. No controlled paediatric trial evidence of functional benefit.
Vojta therapy	Early research only	Reflex locomotion through pressure at defined points. Long history in central Europe, limited controlled evidence, and distress during sessions is a recognised concern.
Conductive education (Pető)	In clinical trials	An education-based, group, whole-day model. Studied more than most; results broadly comparable to other intensive programmes rather than superior.
Patterning (Doman–Delacato)	Not supported by evidence	Repeatedly rejected. Paediatric bodies have advised against it for decades on grounds of absent evidence and heavy family burden.

Sensory therapies

Condition	Evidence	Note
Ayres Sensory Integration (manualised) — participation goals in autism	In clinical trials	Randomised trials exist, using fidelity measures and individualised goal outcomes. Results are promising for functional goals; evidence is not strong enough to call it established.
ASI for core autism features or academic skills	Early research only	A different and weaker claim than improving participation in a specific child.
Weighted vests and blankets for attention or behaviour	Not supported by evidence	Studied repeatedly; benefit for attention has not been demonstrated. Blankets may help some children settle at night, which is a comfort claim, not a therapy claim.
Auditory integration and listening programmes	Not supported by evidence	Reviewed multiple times without demonstrated benefit.
Brushing protocols, swings and sensory diets as stand-alone treatment	Early research only	Widely delivered, largely untested as discrete interventions.

Vision and cerebral visual impairment

Condition	Evidence	Note
Identifying CVI in children with cerebral palsy or brain injury	Established care	Standard of care. CVI is the leading cause of visual impairment in children in high-income countries, and it is substantially under-recognised in cerebral palsy.
Environmental and task adaptation for CVI	Established care	Reducing visual clutter, controlling contrast and lighting, allowing processing time — established practice and often the single most effective change.
Structured visual habilitation programmes	In clinical trials	Programmes to build visual behaviours are widely used; controlled evidence of benefit over adaptation alone is limited.
Correcting refractive error and treating ocular problems	Established care	Basic and frequently overlooked. A child can have CVI and need glasses.

Dietary therapies

Condition	Evidence	Note
GLUT1 transporter deficiency, pyruvate dehydrogenase deficiency	Established care	Here the ketogenic diet is not an adjunct but the treatment itself: it supplies the brain with the fuel it cannot otherwise obtain. This is also why the diagnosis matters — the condition is treatable and will not be found without testing for it.
Drug-resistant epilepsy in children	Established care	Randomised evidence supports the ketogenic diet as an option after antiseizure medicines have failed, with roughly a third to a half of children achieving a meaningful reduction in seizures. It is recommended in international guidance and delivered by a specialist team.
Specific epilepsy syndromes — infantile spasms, Dravet, myoclonic-atonic epilepsy, Lennox-Gastaut, FIRES	Established care	These are the syndromes in which dietary therapy is most often considered early rather than late, and in several of them response rates are higher than in drug-resistant epilepsy generally.
Gluten-free, casein-free diet for autistic features	Not supported by evidence	Systematic reviews have not found adequate evidence of benefit, and the better-designed double-blind studies with controlled food challenges were negative. The separate and important exception is coeliac disease, which must be diagnosed and treated if it is present.
GAPS diet	Not supported by evidence	No clinical trial evidence of any kind. It is among the most restrictive regimens offered to children, and in a child who already eats a narrow range the nutritional risk is substantial.
Elimination and few-foods diets in ADHD	In clinical trials	Some children do respond, and controlled studies of restricted few-foods diets report an effect in a subgroup. The burden on the family is high and dietetic supervision is needed, so this is a considered trial rather than a default.
Removing food additives or sugar to change behaviour	Early research only	Effects reported in controlled studies are small and inconsistent. Sugar in particular has been tested repeatedly in blinded conditions without supporting the belief that it causes hyperactivity.

Layer two

Diagnosis-led treatment

Getting a precise diagnosis

Condition	Evidence	Note
Developmental delay, intellectual disability, epilepsy of unknown cause	Established care	Exome or genome sequencing is recommended as a first-tier test in international guidance, with diagnostic yield that no amount of further imaging matches.
Cerebral palsy without a clear injury on MRI	Established care	A meaningful proportion of children labelled cerebral palsy have an underlying genetic condition. Where the history and MRI do not explain the picture, genetic testing is indicated.
Early-onset epileptic encephalopathy	Established care	Diagnosis frequently changes treatment directly — some genes make particular antiseizure medicines the right choice and others actively harmful.
Neuromuscular presentations	Established care	Newborn screening and early genetic diagnosis in spinal muscular atrophy exist precisely because treatment before symptoms changes the outcome.

Gene-targeted treatments

Condition	Evidence	Note
Spinal muscular atrophy	Established care	Several licensed treatments with different mechanisms: an intrathecal antisense oligonucleotide, a one-off AAV gene replacement, and an oral splicing modifier. This condition was reshaped within a decade.
CLN2 Batten disease	Established care	Enzyme replacement delivered into the brain's ventricles slows functional decline.
Metachromatic leukodystrophy	Established care	An ex vivo gene therapy using the child's own corrected stem cells, licensed in Europe for early stages, before symptoms are established.
AADC deficiency	Established care	A gene therapy delivered directly into the brain, licensed in Europe.
Duchenne muscular dystrophy	In clinical trials	Exon-skipping oligonucleotides and an AAV gene therapy are licensed in some jurisdictions under accelerated pathways; the size of the functional benefit remains debated.
Dravet syndrome, Angelman syndrome, other channelopathies	In clinical trials	Antisense approaches are in clinical trials. Not approved. This is where a registered trial is a reasonable thing to seek out.
Cerebral palsy from acquired injury	Not supported by evidence	There is no gene to target. Genetic treatments have no role where the cause is an injury rather than a gene.

Layer three

Emerging treatments

Stem cell therapy

Condition	Evidence	Note
Blood and immune disorders	Established care	Cord blood transplantation is a licensed, decades-old treatment — this is where the word “approved” genuinely applies.
Cerebral palsy	In clinical trials	Randomised controlled trials have been run and more are under way. Where benefit is reported it is usually small and measured on motor scales, not a return to typical development.
Hypoxic-ischaemic encephalopathy (newborn)	In clinical trials	Early-phase trials, usually added to cooling. Feasibility and safety are the main findings so far.
Autism	Early research only	A well-conducted randomised trial of cord blood did not meet its main endpoint. A possible signal in a subgroup has not been confirmed in a separate trial.
Genetic epilepsies	Early research only	No controlled paediatric evidence. Biological reasoning only.

Exosomes

Condition	Evidence	Note
Any paediatric neurological condition	Early research only	Laboratory and animal work is substantial. Controlled trials in children are not.
Marketed “exosome” products	Not supported by evidence	EU law now allows these to be developed as biological medicines, but no product has yet completed that path: none is authorised anywhere. Preparations sold outside trials are not standardised, and regulators have warned about them after serious infections traced to unlicensed batches.

Muse cells

Condition	Evidence	Note
Newborn hypoxic-ischaemic encephalopathy	Early research only	Early-phase Japanese work exists, given alongside standard therapeutic cooling. The numbers are small and the question being asked is safety and feasibility, but this is a properly conducted clinical programme rather than an informal offering.
Adult stroke, myocardial infarction, ALS, spinal cord injury	In clinical trials	Registered early-phase trials have been conducted with a defined, manufactured product. These are the studies that established the safety profile on which paediatric work now builds.
Cerebral palsy, autism	Early research only	No controlled paediatric trial has reported on these conditions. Individual positive responses are described by treating clinicians and families; those observations are worth recording systematically, and they are not the same thing as a trial result.

Photobiomodulation

Condition	Evidence	Note
Autism	Early research only	Small randomised and sham-controlled studies report behavioural improvements, and the number of positive reports has grown steadily. Samples are modest, follow-up short and devices differ between studies, so the size of the effect is not yet pinned down — but the reports point the same way.
Traumatic brain injury (mostly adults)	Early research only	The largest body of transcranial work, still early and mainly in adults.
Cerebral palsy, epilepsy	Not supported by evidence	No controlled paediatric evidence for seizure control or motor outcome.

Magnetic stimulation

Condition	Evidence	Note
Adolescent depression	Established care	Regulators in some countries have cleared repetitive TMS as an add-on treatment in older adolescents. Approval is specific to that indication, age range and device.
Autism — repetitive behaviours and stereotypies	In clinical trials	This is where the autism signal is most consistent. Low-frequency protocols over the dorsolateral prefrontal cortex have repeatedly reported reductions in repetitive behaviours and stereotypies, alongside changes on attention and event-related potential measures. Much of the work comes from a small number of groups and the trials are modest in size, so it is not yet established care — but the direction has held across studies.
Hemiparetic cerebral palsy — upper limb	In clinical trials	The strongest paediatric motor evidence, and it depends on combination. Randomised work pairing low-frequency rTMS over the uninvolved hemisphere with intensive constraint or bimanual therapy has found the combined arm doing better than either component alone. Read as an amplifier of good hand therapy it has a real case; read as a stand-alone treatment it does not.
Cerebral palsy — gait and lower limb	Early research only	Much less studied than the upper limb, with smaller series and less consistent results.
Epilepsy	Early research only	Low-frequency protocols have been explored for focal epilepsy. Evidence is limited, and stimulation parameters matter for safety.

Peptide preparations

Condition	Evidence	Note
Adult ischaemic stroke (cerebrolysin)	In clinical trials	The most studied use, and genuinely contested. A Cochrane review concludes it probably does not reduce death; other meta-analyses report early neurological improvement. Adults, not children.
Children — developmental delay, cerebral palsy, autism	Early research only	Widely prescribed in parts of Eastern Europe, Russia, Türkiye and Asia. Controlled paediatric evidence is thin and mostly low quality.
“Peptide therapy” from wellness and anti-ageing clinics	Not supported by evidence	BPC-157, thymosin fragments, growth-hormone secretagogues and similar. Not licensed as medicines for children anywhere, no paediatric neurological evidence, and several appear on anti-doping lists.

Medicinal mushrooms and nootropics

Condition	Evidence	Note
Lion's mane for nerve growth — children	Early research only	Compounds in this fungus stimulate nerve growth factor in laboratory models. That is where the evidence stops; there are no paediatric neurological trials.
Citicoline	In clinical trials	The most studied compound in this group, mainly in adult stroke and cognitive decline, with mixed results. Paediatric neurological evidence is limited.
Piracetam and related racetams	Early research only	Long history, particularly in Europe. Evidence for benefit in children is weak and the drug is not licensed for these uses in most countries.
Psilocybin	Not supported by evidence	Adult psychiatric trials are serious science. There is no basis for use in children, and it is not a neurorehabilitation treatment.
General “brain support” supplement blends	Not supported by evidence	Multi-ingredient products marketed for focus, speech or development. No evidence, variable contents, and interactions with antiseizure medicines are rarely considered.

Cannabidiol and cannabis-based products

Condition	Evidence	Note
Dravet syndrome	Established care	Purified cannabidiol is licensed as an add-on antiseizure medicine. In the pivotal trial the median monthly convulsive seizure frequency fell by about 39% on cannabidiol against about 13% on placebo — a real effect, and a partial one.
Lennox-Gastaut syndrome	Established care	Licensed on the same basis, with randomised trials showing a reduction in drop seizures. In the European licence it is approved as an add-on to clobazam; the United States licence does not tie it to clobazam.
Tuberous sclerosis complex	Established care	Added to the licence after a randomised trial in TSC-associated seizures — the FDA in 2020, the European Commission in 2021.
Other drug-resistant epilepsies and developmental epileptic encephalopathies	In clinical trials	A substantial and consistent real-world body of evidence, even though it is not randomised. In the four-year results of the United States expanded access programme, 51–59% of patients had at least a halving of convulsive seizures and 11–17% became seizure-free. Recent multicentre cohorts in developmental and epileptic encephalopathies report roughly half to two thirds reaching a 50% reduction at one to two years. This is class III evidence rather than licence-grade — but it is a long way from nothing.
CDKL5 deficiency, Dup15q, Angelman, Doose, Aicardi, Sturge-Weber, FIRES	In clinical trials	Each has its own open-label or retrospective series, and in most a meaningful proportion of children respond. Two honest caveats: a prospective series in CDKL5 found early responses often faded by twelve months, and in several syndromes families continued the drug for gains in alertness, sleep and behaviour rather than seizure counts alone.
Autism	Early research only	Randomised trials of CBD-rich extracts have been run, including a crossover trial in Israel. Results are mixed and the main outcomes have generally not been met. Not established care.
Spasticity in cerebral palsy	Early research only	Most of the spasticity evidence is from nabiximols in adults with multiple sclerosis. Paediatric cerebral palsy data are thin.
Shop-bought CBD oils for any neurological condition	Not supported by evidence	These are food supplements, not medicines. Independent analyses repeatedly find the amount of cannabidiol differing from the label, and sometimes detectable THC. A dose you cannot verify is a dose you cannot evaluate.

Chelation therapy

Condition	Evidence	Note
Proven heavy metal poisoning (lead, mercury, arsenic)	Established care	Here chelation is real medicine and can be life-saving. It is done in hospital, for a poisoning confirmed by blood levels, with the agent, dose and route chosen by a toxicologist and the child monitored throughout. Everything below concerns something else entirely.
Autism	Not supported by evidence	A Cochrane review found no clinical trial evidence that chelation is an effective treatment for autism. The one randomised study it could include was of poor quality. There is nothing to weigh against the risks.
Cerebral palsy, developmental delay, ADHD	Not supported by evidence	No evidence of any kind. These conditions are not caused by heavy metal accumulation, and chelating a child who is not poisoned removes essential minerals they need to grow.
Treatment based on a “provoked” urine metal test	Not supported by evidence	Giving a chelating agent and then measuring metals in the urine will always raise the result, because that is what the drug does. The laboratory reference ranges printed beside it are for unprovoked samples. A raised provoked result is not a diagnosis of poisoning — it is the predictable output of the test.

Probiotics, prebiotics and the gut–brain axis

Condition	Evidence	Note
Preventing necrotising enterocolitis in very preterm infants	In clinical trials	The strongest evidence anywhere in this field, from many randomised trials and now used routinely in some neonatal units. But see the safety section: the FDA warned in 2023 after a preterm baby died of sepsis caused by the organism in the product given. Strain, product quality and who the baby is all matter.
Antibiotic-associated diarrhoea	Established care	Specific strains reduce it. A bounded, short-term, well-defined use.
Acute infectious gastroenteritis	In clinical trials	Older evidence suggested modest benefit; two large randomised trials in children published in 2018 found no benefit. Enthusiasm has fallen accordingly.
Infant colic	In clinical trials	Lactobacillus reuteri has reasonable evidence in breastfed infants and much weaker evidence in formula-fed infants.
Autism — core features	Early research only	Several small randomised trials, mostly null on their primary outcome. A 2025 meta-analysis pooling 15 trials of mixed gut-directed interventions found a small significant effect (about −0.12), but it mixes probiotics, prebiotics, diet and faecal transfer, so it says little about any one of them.
Autism — gut symptoms	In clinical trials	The more reasonable target. Constipation and discomfort are common and genuinely worth treating, with probiotics as one modest option among several.
Faecal microbiota transfer	Early research only	The widely shared autism results came from a small open-label study with no control group. That design cannot separate treatment from natural change and expectation. Randomised trials are under way; results are not in.
Constipation in cerebral palsy	Early research only	Common and important, but standard management — fluids, fibre, laxatives, positioning, reviewing medications — has far better support than probiotics.

How to use this honestly

A row marked “early research only” is not a verdict that the treatment is useless. It is a statement that nobody can yet tell you whether it will help your child — which is exactly the information a family needs before committing time and hope to it.