Can restoring the brain's waste-clearance system improve brain function in Alzheimer’s and some autism?

 

One of the most exciting developments in modern neuroscience is the growing understanding of the glymphatic system—the brain's own waste-clearance system.

This network circulates cerebrospinal fluid (CSF) through the brain, removing metabolic waste products and helping maintain a healthy neuronal environment. Remarkably, the glymphatic system is most active during deep (slow-wave) sleep, suggesting that one of sleep's most important functions is to clean the brain.

Interest in this system has exploded over the past decade because impaired glymphatic function has now been implicated in Alzheimer's disease, Parkinson's disease, traumatic brain injury, stroke and several other neurological disorders.

Increasingly, researchers are also asking whether glymphatic dysfunction contributes to autism spectrum disorder (ASD).

 

A remarkably recent discovery

One of the most surprising aspects of this story is just how new it is.

The glymphatic system was only discovered in 2012 by Professor Maiken Nedergaard and colleagues. Until then, neuroscientists knew that the brain produced metabolic waste, but nobody really understood how it was removed. Unlike the rest of the body, the brain appeared to have no conventional lymphatic system.

The discovery of the glymphatic system transformed our understanding of brain biology. Researchers showed that cerebrospinal fluid flows alongside arteries into the brain, exchanges with the fluid surrounding brain cells, and then carries away waste products before leaving alongside veins.

Only three years later, in 2015, another major breakthrough followed when researchers discovered meningeal lymphatic vessels surrounding the brain. These vessels drain fluid from the glymphatic system into the body's lymphatic system.

Together, these discoveries revealed that the brain possesses its own sophisticated waste-disposal network.

Considering that this entire field is little more than a decade old, it is remarkable how quickly it has expanded. Today, impaired glymphatic function has been linked to Alzheimer's disease, Parkinson's disease, multiple sclerosis, traumatic brain injury, stroke, depression, schizophrenia and, increasingly, autism.

 

Autism and the glymphatic system

Until recently, the idea that glymphatic dysfunction might contribute to autism was based largely on indirect observations.

Researchers had reported:

• enlarged extra-axial cerebrospinal fluid in some infants who later develop autism,
• enlarged perivascular spaces,
• chronic neuroinflammation,
• and the very high prevalence of sleep disorders in autism.

Now, neuroimaging studies are beginning to provide more direct evidence.

A newly published MRI study examined 78 children with autism and 48 typically developing controls using the DTI-ALPS technique, a non-invasive MRI method that estimates glymphatic activity. The investigators found significantly reduced glymphatic function in the autism group. Even more interestingly, poorer glymphatic function correlated with more severe communication difficulties and poorer visual-motor integration. Statistical modelling suggested that impaired visual-motor integration partly explained the relationship between glymphatic dysfunction and communication deficits.

Glymphatic system dysfunction in children with autism spectrum disorder as evidenced by the diffusion tensor imaging along perivascular spaces index

The authors concluded that impaired glymphatic clearance may contribute to the pathophysiology of autism and represents a promising area for future mechanistic and interventional research.

These studies do not prove that impaired glymphatic clearance causes autism. However, they do suggest that it may contribute to the biology of at least a subgroup of autistic individuals.

 

Why deep sleep matters

Perhaps the most important aspect of the glymphatic system is that it works best during deep slow-wave sleep.

During deep sleep:

• cerebrospinal fluid flows more efficiently through the brain,
• the space between brain cells expands,
• metabolic waste products are removed,
• inflammatory molecules are cleared,
• and the brain performs what is essentially its nightly housekeeping.

Sleep disturbances affect around 50–80% of autistic children, often continuing into adulthood. If glymphatic clearance depends on deep sleep, this raises the possibility that chronic sleep disruption contributes to impaired waste clearance, neuroinflammation and altered brain function, creating a vicious cycle.

This may be one reason why improving sleep often has benefits that extend well beyond simply reducing daytime fatigue.

 

Aquaporin-4: the brain's plumbing protein

The glymphatic system depends on a protein called aquaporin-4 (AQP4), which forms tiny water channels in the end-feet of astrocytes—support cells that surround the brain's blood vessels.

These channels allow cerebrospinal fluid to move efficiently between blood vessels and brain tissue, helping wash away metabolic waste products. In animal experiments, deleting AQP4 dramatically reduces glymphatic clearance, highlighting its central role in the brain's waste-disposal system.

Interestingly, several studies have reported altered AQP4 expression or localisation in autism. Although it is still unclear whether these changes are a cause or a consequence of autism, they provide another possible explanation for why glymphatic function may be impaired in at least some autistic individuals.

 

Exercise may also support the brain's cleaning system

Exercise is another intervention that may enhance glymphatic function.

Experimental studies suggest that regular physical activity can:

• improve vascular health,
• improve sleep quality,
• enhance cerebrospinal fluid dynamics,
• reduce neuroinflammation,
• and potentially improve glymphatic clearance.

Even if future studies show only modest effects on glymphatic function, exercise already has well-established benefits for cardiovascular health, cognition, mood and healthy ageing.

 

 

 A fascinating new therapeutic idea

This brings us to one of the most intriguing studies published in the last 12 months.

The influence of intermittent hypercapnia on cerebrospinal fluid flow and clearance in Parkinson’s disease and healthy older adults

Researchers investigated whether they could artificially stimulate the glymphatic system by having volunteers breathe alternating short periods of air containing 5% carbon dioxide, followed by normal air.

Rather than continuously increasing carbon dioxide, the intermittent exposure generated rhythmic expansion and contraction of cerebral blood vessels.

Brain imaging demonstrated increased cerebrospinal fluid movement, while blood tests suggested increased movement of brain-derived proteins into the circulation, consistent with enhanced glymphatic clearance.

Importantly, this was not an autism or Alzheimer's treatment study. It was a proof-of-concept experiment involving healthy older adults and people with Parkinson's disease.

Nevertheless, it demonstrated something remarkable:

The human glymphatic system appears to be modifiable.

Mimicking what happens during deep sleep

Perhaps the most fascinating aspect of the study is the proposed mechanism.

The intermittent carbon dioxide exposure generated slow vascular oscillations remarkably similar to those naturally seen during deep sleep, when glymphatic clearance reaches its maximum.

In effect, the researchers may have temporarily reproduced one of the physiological mechanisms by which deep sleep naturally cleans the brain.

That opens an entirely new therapeutic concept.

Rather than trying to remove individual proteins such as amyloid with drugs, perhaps we can improve the brain's own housekeeping system.

 

How the intermittent CO₂ therapy was performed

The researchers did not expose participants to continuous carbon dioxide. Instead, they used a carefully controlled pattern of intermittent increase in the level of carbon dioxide (CO₂) in the blood (hypercapnia). The hypercapnia was designed to create rhythmic changes in cerebral blood flow.

Interestingly, several studies have reported abnormalities in cerebral blood flow regulation in autism, including reduced blood flow in specific brain regions and altered responses of cerebral blood vessels to changing carbon dioxide levels. Since the glymphatic system depends on rhythmic vascular pulsations to drive cerebrospinal fluid through the brain, impaired vascular regulation could potentially contribute to reduced glymphatic clearance.

Participants underwent three treatment sessions, each lasting approximately 10 minutes, for a total treatment time of about 30 minutes.

During each session they repeatedly alternated between:

• 35 seconds breathing a gas mixture containing 5% carbon dioxide
• 35 seconds breathing normal air

This cycle was repeated 24 times across the three sessions. The sessions were performed during a single visit, with short breaks between them.

The rationale was that the repeated dilation and relaxation of cerebral blood vessels would generate slow vascular oscillations similar to those that naturally occur during deep slow-wave sleep, when the brain's glymphatic system is most active.

MRI scans demonstrated increased cerebrospinal fluid movement through the brain, while blood tests showed increased concentrations of several brain-derived proteins—including amyloid-β, tau, GFAP and neurofilament light—in the circulation. The researchers interpreted these findings as evidence of enhanced glymphatic clearance rather than brain injury.

 

A new way of thinking

Much autism research has focused on inflammation, oxidative stress, mitochondrial dysfunction and altered neurotransmission.

But perhaps, in some autistic individuals, these abnormalities are partly downstream consequences of impaired brain waste clearance.

Instead of asking:

Which drug treats autism?

perhaps we should also ask:

Can we restore the brain's own housekeeping system?

That strategy might include:

• protecting deep sleep
• treating sleep disorders
• exercising regularly
• maintaining good cardiovascular health
• and perhaps one day therapies that directly stimulate glymphatic flow

 

Looking ahead

This remains a hypothesis.

Neither the imaging studies nor the intermittent hypercapnia study prove that improving glymphatic function will improve autism symptoms.

However, taken together they suggest something genuinely exciting.

For the first time we now have:

• evidence that glymphatic dysfunction is present in autism,
• evidence that poorer glymphatic function is associated with more severe communication difficulties,
• and evidence that the human glymphatic system may be therapeutically stimulated.

That does not yet constitute a treatment.

But it does provide the foundations for an entirely new direction of autism research—one aimed not at treating individual downstream abnormalities, but at restoring one of the brain's most fundamental maintenance systems.

If this hypothesis proves correct, the simplest interventions may remain among the most important: protect deep sleep, exercise regularly, maintain good cardiovascular health, and treat sleep disorders whenever possible. Future research will determine whether directly stimulating glymphatic flow can add to these fundamental approaches.

While a variety of therapies are marketed as improving lymphatic drainage, there is currently no convincing evidence that they enhance glymphatic clearance in humans.

Edging closer to targeting neuro-inflammation in autism via FPR2

 

 A thoughtful mouse in Naples

For many years, autism research has focused primarily on neurotransmitters such as GABA and glutamate. More recently, however, increasing attention has turned to another major contributor to brain function: the immune system.

Numerous studies have reported evidence of chronic microglial activation and elevated inflammatory cytokines in at least a subset of autistic people. The obvious question has always been:-

Can reducing neuro-inflammation improve the core features of autism?

A newly published study from researchers at the University of Naples suggests we may be edging closer to answering that question.

Rather than simply suppressing inflammation, they have taken a much more sophisticated approach by activating one of the body's own inflammation-resolution pathways. Their target is a receptor called Formyl Peptide Receptor 2 (FPR2), which has not been covered in previous posts.

 

Agonist MR-39 Supports Synaptic Health in the BTBR Mouse and Features a Favorable Safety Profile

Chronic unresolved inflammation is a common feature of several Central Nervous System (CNS) disorders, including autism spectrum disorder (ASD). We previously demonstrated that Formyl Peptide Receptor 2 (FPR2) activation by our agonist MR-39 reduced several inflammatory markers and improved social behavior in two validated animal models of ASD. Therefore, we decided to delve deeper into the potential of MR-39 as a drug for treating ASD. We first investigated the molecular mechanisms underlying the beneficial effects of MR-39 in BTBR mice. MR-39 significantly normalized pro-inflammatory cytokine release and NF-κB expression in the hippocampus and cortex, resulting in upregulation of synaptophysin protein levels, which, in turn, promote plasticity and correct abnormalities in dendritic spine morphology. Next, we characterized the safety and pharmacokinetic profile of MR-39 with respect to potential advancement for further pre- and clinical studies. We found that MR-39 was not genotoxic and safe to use since it had limited interaction with the majority of the targets associated with the adverse drug reaction. Consistently, a repeated-dose administration study evidenced no clinical signs attributable to treatment-related toxicity. On the other hand, MR-39 exhibited rapid hepatocyte clearance and interaction with efflux systems in vitro, suggesting possible limitations due to its pharmacokinetic properties. Finally, we explored multiple strategies to overcome MR-39’s low aqueous solubility, finding that the cosolvent approach can greatly enhance solubility and wettability. Overall, our study confirmed that promoting inflammation resolution with MR-39 can open new therapeutic options for ASD and that this compound has potential as a drug.

 

Why FPR2 is different

Most anti-inflammatory therapies work by blocking inflammatory pathways.

For example:

• statins reduce inflammatory signalling through multiple mechanisms
• pioglitazone shifts microglia towards a more reparative state via PPAR-γ
• ibudilast suppresses activated microglia
• low-dose naltrexone appears to reduce chronic glial activation through TLR4
• clemastine may reduce microglial activation while simultaneously promoting remyelination.

These are all potentially useful approaches, but fundamentally they are trying to dampen inflammation.

FPR2 works differently.

Inflammation is not simply switched on and then allowed to fade away. The body possesses an active programme that tells the immune system when the danger has passed and it is time to stop fighting and begin repairing damaged tissue.

FPR2 is one of the master regulators of this resolution of inflammation.

This distinction may prove extremely important in neurological disorders where persistent inflammation itself may be preventing normal synaptic development and plasticity.

 

Where is FPR2 found?

One reason FPR2 has attracted so much attention is that it is expressed in many of the cells involved in both inflammation and tissue repair.

Within the brain, FPR2 is found on:

• microglia - the brain's resident immune cells that monitor the environment, remove debris and regulate inflammation
• astrocytes - which support neurons, maintain the blood-brain barrier and participate in immune signalling
• neurons - suggesting FPR2 may influence synaptic plasticity and neuronal survival directly as well as indirectly through the immune system
• brain endothelial cells - which regulate communication between the bloodstream and the brain

Outside the brain it is expressed on many components of the innate immune system including neutrophils, macrophages, monocytes and several other immune cell types.

This broad distribution explains why FPR2 is now being investigated in such diverse conditions as Alzheimer's disease, stroke, multiple sclerosis, inflammatory bowel disease, rheumatoid arthritis, myocardial infarction and now autism.

Unlike many receptors that are confined to a single tissue, FPR2 appears to function as one of the body's master regulators of inflammation resolution. It helps coordinate the transition from an active immune response to tissue repair and restoration of normal homeostasis.

For autism this is particularly interesting because activating FPR2 could simultaneously calm activated microglia, reduce inflammatory cytokines, improve blood-brain barrier function and create an environment in which neurons and synapses can remodel more effectively.

Its widespread expression suggests that FPR2 is not simply controlling inflammation—it is coordinating the body's transition from defence to repair.

 

The experimental drug MR-39

The researchers tested a small molecule called MR-39, an activator of FPR2.

Interestingly, MR-39 was not originally developed as an autism therapy.

It emerged from research into neuro-inflammatory disorders, particularly Alzheimer's disease, illustrating an increasingly important trend in autism research: extending therapies developed for one neurological disorder into another when they target shared biological mechanisms.

Rather than developing drugs specifically "for autism", researchers are increasingly asking:

Which biological pathways are abnormal in autism, and are there drugs already being developed for those pathways?

This strategy has already given us therapies such as bumetanide, metformin, pioglitazone, statins and calcium-channel blockers.

MR-39 is another example.

An additional reason this approach is particularly attractive is that several studies have reported reduced circulating levels of Lipoxin A4 (LXA₄) in autistic children. LXA₄ is not just another anti-inflammatory molecule—it is one of the body's own specialised pro-resolving mediators (SPMs) and one of the principal activators of FPR2, already present in your body.

 

Decreased plasma levels of lipoxin A4 in children with autism spectrum disorders

The aim of this study was to evaluate the plasma levels of lipoxin A4 (LXA4), a mediator involved in the resolution of inflammation in Chinese children with autism spectrum disorders (ASD). From January 2013 to June 2014, a total of 150 children (75 confirmed ASD cases and 75 their age-matched and sex-matched control cases) participated in this study after consent was obtained from their parents. Clinical information was collected. Plasma levels of LXA4 were measured at baseline. The severity of ASD was assessed at admission using the Childhood Autism Rating Scale total score. The results indicated that the mean plasma levels of LXA4 were significantly lower in autistic children compared with the normal children (P<0.0001). There was a significant negative relationship between circulating LXA4 levels and severity of autism evaluated by Childhood Autism Rating Scale scores (P=0.006) after adjustment for the possible covariates.

These results suggested that autistic children had lower plasma LXA4 levels, suggesting an increased susceptibility to recurring inflammation in these samples.

 

In other words, the body already possesses a natural mechanism for activating this receptor and switching inflammation into its resolution phase.

If LXA₄ levels are indeed reduced in at least some autistic individuals, that natural "stop fighting, start repairing" signal may be weakened. MR-39 can therefore be viewed not as introducing an entirely artificial pathway, but as pharmacologically replacing or amplifying a biological signal that may already be deficient.

This provides a much stronger biological rationale for targeting FPR2 in autism than simply identifying it as another interesting receptor.

 

What did the new Italian study find?

Using the well-established BTBR mouse model of autism, the researchers found that after only eight days of treatment MR-39:

• significantly reduced the inflammatory cytokines IL-1β and TNF-α
• normalised NF-κB signalling
• increased levels of synaptophysin, an important marker of healthy synapses
• corrected abnormalities in dendritic spine morphology
• built upon previous work from the same laboratory showing improvements in social behaviour

One particularly encouraging finding was that MR-39 did not simply increase the number of synapses.

Instead, it appeared to improve their maturity.

The abnormal dendritic spines seen in the BTBR mice became much more like those found in healthy animals.

This is important because many forms of autism are characterised not simply by having too many or too few synapses, but by synapses that have failed to mature normally.

The findings therefore suggest that reducing chronic neuro-inflammation may allow existing neural connections to complete a more normal developmental programme, rather than simply creating new ones.

This may explain why MR-39 improved synaptic proteins and dendritic spine morphology without dramatically altering synapse number. Rather than forcing neurons to form new connections, the drug appears to create the biological conditions that allow normal developmental and repair processes to resume.

 

More than another mouse study

Many animal studies simply report behavioural improvements.

This paper goes considerably further.

The authors also examined:

• pharmacokinetics
• liver toxicity
• cardiac safety
• off-target effects
• genotoxicity
• metabolism
• formulation chemistry

In other words, they were already thinking like drug developers rather than purely academic scientists.

That makes this one of the more substantial preclinical autism studies published in recent years.

 

The challenges

MR-39 is not yet ready for human trials.

The authors identified several important limitations including poor oral bioavailability, rapid metabolism and potential cardiac and liver toxicity.

Fortunately, none of these appear insurmountable, and the compound was well tolerated in mice at therapeutic doses.

Like many first-generation research compounds, MR-39 itself may never become the final medicine.

Its greatest contribution may simply be demonstrating that FPR2 is a worthwhile therapeutic target.

 

Autism as a network disorder

One theme that has appeared repeatedly on this blog is that autism is probably best viewed not as a single biochemical defect, but as a network disorder.

Genes, mitochondria, immune activation, metabolism, ion channels, neurotransmitters and synaptic plasticity all interact to produce a stable pattern of brain function. Once established, that pattern may become a new homeostasis—a stable but abnormal equilibrium.

This helps explain why there is unlikely ever to be a single "autism drug".

A calcium-channel blocker may improve neuronal excitability.

A statin may reduce neuro-inflammation.

Pioglitazone may reprogramme microglia.

Clemastine may promote remyelination.

Bumetanide may restore inhibitory signalling.

Each addresses one part of the network.

The hope is not that one intervention completely resets the brain, but that multiple interventions gradually shift the network towards a healthier and more stable state.

Viewing autism as a network disorder naturally leads to a different therapeutic philosophy.

Instead of expecting one drug to correct every abnormality, the aim becomes rational polytherapy—combining carefully selected interventions that each influence a different component of the biological network.

One treatment might reduce neuro-inflammation.

Another might improve mitochondrial metabolism.

Another might restore inhibitory neurotransmission.

Another might promote remyelination.

Another might improve synaptic plasticity.

Individually these therapies may each produce only modest benefits.

Together, however, they may shift the entire network towards a healthier and more stable equilibrium.

This approach is already becoming familiar in other areas of medicine. Multiple sclerosis (MS) provides an excellent example. Twenty years ago, the primary goal of treatment was to suppress immune attack on myelin. Today, researchers increasingly recognise that long-term success also requires promoting remyelination, protecting neurons and encouraging the brain's own repair mechanisms. As a result, therapies such as clemastine, specialised pro-resolving mediators, neuroprotective agents and regenerative approaches are being investigated alongside traditional immunomodulatory drugs.

The philosophy is changing from simply stopping damage to actively promoting recovery.

I believe autism research is beginning to move in the same direction.

FPR2 is particularly interesting because it sits relatively high in the hierarchy of inflammatory regulation. Rather than blocking one inflammatory pathway, it appears to activate one of the brain's own programmes for restoring homeostasis. As such, future FPR2 activators/agonists may eventually become one component of a broader polytherapy approach that aims not merely to suppress symptoms, but to help the brain repair and reorganise itself.

  

Linking FPR2 with the Cell Danger Response

One striking aspect of this work is how well it fits with Robert Naviaux's Cell Danger Response (CDR) hypothesis.

Naviaux proposes that when cells encounter infection, toxins, trauma or metabolic stress they switch into a protective emergency programme.

Mitochondria alter their metabolism.

ATP is released outside cells as a danger signal.

Inflammatory pathways become activated.

Normal cellular communication becomes secondary to survival.

This response is entirely normal.

The problem arises if cells fail to exit this emergency state after the original danger has passed.

Viewed in this way, FPR2 could represent one of the mechanisms that helps cells transition out of the Cell Danger Response and back towards normal homeostasis.

Although this link remains speculative, it provides an intriguing framework that connects mitochondrial signalling, neuroinflammation and synaptic repair.

Interestingly, the two fields approach the problem from opposite directions.

Naviaux focuses on danger signalling—how cells detect injury and enter an emergency programme.

The FPR2 researchers focus on resolution signalling—how the immune system recognises that the danger has passed and initiates tissue repair.

These may simply represent different stages of the same biological programme.

One explains how the emergency begins.

The other may explain how it ends.

Ultimately, both point towards the same biological transition:

 

Stop defending. Start repairing.

Rather than directly repairing neurons, FPR2 activation may simply remove one of the major barriers preventing the brain from repairing itself.

 

Helping the body repair itself

Perhaps the most interesting lesson from this paper extends well beyond autism.

For decades drug development has largely focused on blocking abnormal pathways.

Block an enzyme

Block a receptor

Block a cytokine

Increasingly, medicine appears to be moving towards a different philosophy.

Many of the body's repair mechanisms already exist.

Stem cells migrate to damaged tissue.

Microglia clear cellular debris.

Specialised lipid mediators resolve inflammation.

Neurons remodel synaptic connections.

Oligodendrocytes repair myelin.

The challenge is often not that these systems are absent, but that they have become stalled or trapped in an abnormal steady state.

FPR2 appears to be one of the molecular switches that tells the immune system:

"The danger has passed. Stop fighting. Start rebuilding."

That may explain why FPR2 is attracting attention not only in autism, but also in Alzheimer's disease, stroke, multiple sclerosis and cardiovascular disease.

Rather than overriding biology, it appears to encourage biology to resume doing what evolution designed it to do.

 

Can we simply increase Lipoxin A4 instead?

An obvious question is: if autistic children have reduced levels of Lipoxin, why not simply treat with Lipoxin? Unfortunately, this is not practical. Like many of the body's own signalling molecules, Lipoxin is chemically unstable and has a biological half-life measured in minutes. It is rapidly broken down after being produced, making it a poor drug candidate.

This is the main reason researchers focus on developing stable FPR2 activators like MR-39, which are engineered to mimic the beneficial effects of Lipoxin while remaining active in the body for much longer. 

Another strategy is to encourage the body to produce more of its own Lipoxin. Unlike the resolvins, which are derived from omega-3 fatty acids, Lipoxin is synthesised from arachidonic acid (AA)—an omega-6 fatty acid that is already abundant in most people's cell membranes. The limiting factor here is not the availability of raw materials, but the activity of the specific enzymes required to convert that arachidonic acid into Lipoxin. At present, no supplement has been convincingly shown to raise Lipoxin levels in humans. 

However, there is growing evidence that regular aerobic exercise promotes the production of the body's broader family of specialised pro-resolving mediators (SPMs). For now, until stable analogues or molecular activators clear clinical trials, optimizing these natural physiological pathways remains our best practical tool.

 

Conclusion

MR-39 itself may never become an approved medicine. What encourages me far more is the direction in which neuroscience is moving. Research is gradually shifting away from searching for a single defective molecule and towards understanding the brain as a dynamic biological network capable of adaptation, repair and recovery.

The same trend is appearing across medicine. Many of today's most exciting therapies no longer attempt to repair the body themselves. Instead, they help the body repair itself. Whether through inflammation-resolution pathways such as FPR2, remyelination with clemastine, metabolic reprogramming with pioglitazone, restoration of inhibitory signalling with bumetanide, or perhaps one day therapies based on the Cell Danger Response, the underlying philosophy is becoming remarkably similar. The objective is not simply to suppress disease. It is to restore the conditions under which the brain can heal itself.

Increasingly, researchers are discovering that many chronic diseases are characterized not simply by excessive inflammation, but by a failure of that inflammation to resolve properly. Rather than continually developing stronger anti-inflammatory drugs to stomp out symptoms, the future lies in identifying the biological signals that tell the brain it is finally safe to stop defending itself and start developing normally again. If that philosophy proves correct, future autism treatment may consist of carefully selected combinations of therapies, each nudging a different part of the network in the same direction. Individually they may produce modest improvements, but together they may help the brain escape an abnormal equilibrium and settle into a healthier one.

In my specific case, I think we have already achieved this with my son’s Polypill therapy.

Overcoming Picky Eating and ARFID: What the Latest Research Tells Parents

 

  

A few days ago I read a comment from the parent of an autistic teenager that perfectly illustrates why we should never assume that a restricted diet is permanent.

Six years ago, their son ate only a handful of foods. Every meal was a battle. Introducing anything new seemed impossible. Family meals revolved around avoiding conflict, and eating outside the home was stressful for everyone.

Today, that same young man eats what most people would consider a perfectly normal adult diet. Vegetables, fish, different cuisines and healthy foods that once seemed unimaginable are now part of everyday life.

Nothing miraculous happened.

There was no breakthrough drug.

There was no secret supplement.

There were simply six years of patient, structured work by parents who refused to believe that their child's beige diet was fixed forever.

The journey was not easy. There were setbacks, disappointments and many failed attempts. Progress was measured in months and years rather than days and weeks.

But they never stopped trying.

Their story reminds us of something that is easy to forget. Today's diet does not have to be tomorrow's diet

That message is particularly timely because a major randomized clinical trial from Stanford University has just provided the strongest evidence yet that parents themselves can play a central role in helping children with ARFID make meaningful progress.

 

Three Messages I Hope Every Parent Remembers

If you read nothing else in this article, I hope you remember these three ideas.

 

First, ARFID is an observation, not an explanation.

It describes a child's eating behaviour, but it does not explain why that behaviour exists.

 

Second, ARFID is a diagnosis, not a prognosis.

Receiving the diagnosis tells you where your child is today. It says very little about where they could be five years from now.

 

Third, parents are not passive observers.

The strongest clinical evidence we now have suggests that parents are one of the most important parts of the treatment.

These three ideas underpin everything that follows.

 

ARFID Is a Diagnosis, Not a Prognosis

Avoidant Restrictive Food Intake Disorder (ARFID) was only added to the Diagnostic and Statistical Manual (DSM-5) in 2013. Before then, many children were simply described as "extremely picky eaters."

The diagnosis has been helpful because it acknowledges that severe food restriction is a genuine medical and psychological problem rather than simply bad behaviour or poor parenting.

However, diagnoses can sometimes have unintended consequences.

Some parents hear the word ARFID and begin to think that their child's eating habits are largely fixed.

That is understandable, but it is not what the diagnosis means.

ARFID tells us that eating has become sufficiently restricted to affect health, growth or everyday life.

It does not tell us why.

Nor does it tell us what the future holds.

In medicine we often confuse diagnosis with prognosis.

The two are completely different.

A diagnosis describes today's problem.

A prognosis attempts to predict tomorrow.

The Stanford study—and many individual family experiences—suggest that today's eating habits should not be viewed as a reliable predictor of where a child may be after several years of appropriate intervention.

 

Why Expectations Matter

One lesson I have learned from writing this blog is that expectations matter.

Not because optimism magically changes biology, but because expectations influence how much effort people are prepared to invest.

Autism provides many examples.

Poor handwriting is extremely common. Motor planning, muscle control and coordination are often affected.

Yet many autistic children spend years practising handwriting and eventually develop neat, legible writing. The neurological differences have not disappeared. The skill has improved.

Toilet training provides another example.

Some autistic children remain in diapers/nappies or pull-ups for years because everyone assumes they simply are not ready.

Other families invest months—or sometimes years—using structured toilet-training programs.

Not every child achieves complete independence or perfect handwriting.

But many achieve far more than anyone initially thought possible.

Eating should be viewed in exactly the same way.

It is another developmental skill.

Some children acquire it naturally.

Others require hundreds or even thousands of opportunities to practise.

Of course, families differ enormously.

Parents working full-time, caring for several children or home-schooling may genuinely struggle to devote the time required for intensive feeding programmes.

Those constraints are real and deserve understanding.

However, where circumstances allow, an ARFID diagnosis should encourage parents to increase their efforts—not reduce their expectations.

The diagnosis should increase expectations for intervention, not lower expectations for progress.

 

How Common Are Picky Eating and ARFID?

Picky eating is almost a normal part of childhood.

Around one quarter of young children go through a period when they refuse many foods. Fortunately, most gradually grow out of it.

Autism is different. Depending on the study, between 46% and nearly 90% of autistic children show significant food selectivity.

For some children this simply means having a short list of preferred foods.

For others, eating becomes so restricted that nutritional deficiencies develop, weight falters or everyday family life becomes dominated by food.

This is where ARFID begins.

Rather than being a completely separate condition, it is often helpful to think of ARFID as representing the severe end of a spectrum.

Normal childhood picky eating lies at one end.

Severe nutritional compromise lies at the other.

 

ARFID Is an Observation, Not an Explanation

Perhaps the most important question parents should ask is not:

"Does my child have ARFID?"

Instead ask:

"Why does my child have ARFID?"

The diagnosis simply tells us what is happening.

It does not explain why.

In autistic children there are often multiple contributing factors.

Some children have genuine sensory hypersensitivity.

Textures that seem perfectly ordinary to us may feel intensely unpleasant to them.

Others have chronic gastrointestinal discomfort.

Reflux, constipation, delayed stomach emptying and eosinophilic esophagitis can all make eating uncomfortable.

If every meal is associated with discomfort, avoiding food becomes entirely understandable.

Children with connective tissue disorders such as hypermobile Ehlers-Danlos syndrome, or milder extracellular matrix abnormalities, may also develop gastrointestinal dysmotility, reflux and abdominal pain. These conditions are increasingly recognised in autism and may contribute to restricted eating in a subgroup of children.

Oral-motor difficulties are another overlooked cause.

Some children struggle to chew particular textures efficiently, making certain foods genuinely difficult rather than simply disliked.

Anxiety also plays an important role.

A frightening choking episode or severe vomiting illness can lead to persistent fear of eating.

Finally, nutritional deficiencies themselves may worsen the problem.

Iron deficiency, zinc deficiency and other micronutrient deficiencies can alter taste perception, appetite and energy levels, creating a vicious cycle in which poor diet perpetuates itself.

The important message is this:

Finding one of these biological problems does not mean behavioural therapy is unnecessary.

It means behavioural therapy is more likely to succeed once the underlying problem is treated.

Medical treatment and feeding therapy should not be viewed as competing approaches.

In many children they complement one another.

 

Has modern food made ARFID more common?

One question that is rarely discussed is whether modern food itself may unintentionally reinforce food selectivity.

Many processed foods are engineered to be identical every single time.

Every chip/crisp tastes the same.

Every chicken nugget has the same texture.

Every biscuit/cookie feels identical.

Fresh food is completely different.

One strawberry is sweeter than the next.

One apple is crisp while another is soft.

Even two bananas picked from the same bunch taste slightly different.

For children who crave predictability, processed foods offer exactly that.

Nature does not.

This raises an interesting possibility.

Could a highly processed diet make it even harder for some children to accept the natural variability of real food?

We do not yet know the answer.

But it is certainly an area worthy of research.

 

Why dietary diversity matters

The goal of feeding therapy is not simply to make the list of accepted foods longer.

The goal is to improve health.

A child who expands their diet from five processed beige foods to ten processed beige foods has certainly made progress—but probably not enough.

The greatest benefits come from gradually introducing foods that provide nutrients missing from the existing diet.

Vegetables.

Fruit.

Fish.

Legumes.

Nuts.

Seeds.

Fermented foods.

Whole grains.

Each contributes something different.

Dietary diversity also feeds the gut microbiome.

Different bacteria thrive on different fibres and plant compounds. A monotonous diet supports a relatively monotonous microbiome.

Every additional plant food potentially feeds different bacterial species.

Given the growing evidence linking the gut microbiome to immune function, gastrointestinal health and possibly even brain function, this may become one of the most important long-term benefits of improving diet.

For autistic children, a broader diet may improve growth, bone health, immune function, gastrointestinal health and reduce nutritional deficiencies.

The goal is not simply to produce a child who eats more foods.

The goal is to produce a healthier child.

 

When does ARFID become dangerous?

Not every child who is a picky eater has ARFID.

Many children survive for years on a limited selection of foods and continue to grow normally. Although parents understandably worry, these children often improve naturally as they get older.

ARFID becomes a medical disorder when food restriction begins to cause significant problems. These may include:

• Poor weight gain or weight loss
• Slowed growth
• Nutritional deficiencies (iron, zinc, vitamin D, vitamin C and others)
• Dependence on nutritional supplements or tube feeding
• Extreme anxiety surrounding meals
• Family life becoming dominated by food

The distinction is important because the goal is not to pathologize every fussy eater. It is to identify children whose restricted eating is affecting their health or development.

Fortunately, even severe ARFID is treatable.

 

The Stanford Randomized Trial

The best evidence to date comes from researchers at Stanford University, who recently completed the first large randomized controlled trial of a treatment called ARFID Parent Training Protocol (ARFID-PTP).

Rather than providing months of intensive therapy directly to the child, the researchers trained parents.

This is a subtle but important shift.

Instead of trying to change the child during a one-hour therapy session each week, parents learn how to create hundreds of learning opportunities during normal family life.

The study involved 105 children aged 5–12 years with ARFID.

Families were randomly assigned either to receive the parent-training programme or to continue with usual care.

After six months, the differences were striking.

Children whose parents received the training accepted significantly more new foods, had fewer ARFID symptoms and were more likely to no longer meet diagnostic criteria for ARFID.

Perhaps even more impressive was that parents themselves became more confident and less anxious about feeding.

The therapy had changed not only children's behaviour but also the behaviour of the adults supporting them.

That may be one reason it worked so well.

 

The Therapist Becomes the Coach

Traditionally we imagine therapy as something that happens inside a clinic.

A therapist works with a child while parents wait outside.

Feeding therapy is different.

The therapist's real job is often to coach the parents.

Parents are present at breakfast.

Parents are present at lunch.

Parents are present at dinner.

That means they have thousands of opportunities each year to reinforce progress.

A therapist may only have fifty hours with a child over an entire year.

Parents may have over one thousand mealtimes.

Once parents understand the principles, they become the treatment.

The Stanford study confirms what many experienced feeding therapists have believed for years: empowering parents may be one of the most effective interventions available. 

Two treatments help ARFID, a common pediatric eating disorder, Stanford Medicine trial shows

Family vs Individual Treatment for Children With Avoidant/Restrictive Food Intake Disorder: A Randomized Clinical Trial

To examine the comparative efficacy of Family-based Treatment for Avoidant/Restrictive Food Intake Disorder (FBT-ARFID) to individual Psychoeducational Motivational Therapy (PMT) for underweight children with ARFID between the ages of 6 and 12 years of age. The main outcome evaluated was the difference between groups on change in percent estimated body weight (%EBW) from baseline (BL) to end of treatment (EOT).

Method

Ninety-eight children with ARFID were randomized to 14 sessions over 4 months of telehealth FBT-ARFID or PMT. Assessments of weight/height, eating-related cognitions, and behaviors associated with ARFID were collected online at BL, 1 month, 2 months, and EOT by assessors masked to treatment condition.

Results

FBT-ARFID was superior to PMT at the EOT in promoting increased %EBW. There were no differences between groups on improvements in overall severity of ARFID symptoms or other related ARFID symptoms; however, BL severity of ARFID symptoms moderated the effect, with children who were most symptomatic improving significantly more in FBT-ARFID than in PMT (exploratory analyses).

Conclusion

FBT-ARFID is superior to PMT for promoting weight gain in low-weight children with ARFID, especially for those children with greater severity of ARFID symptoms.

 

 

What Feeding Therapy Actually Involves

Many people imagine feeding therapy consists of persuading a child to eat vegetables.

In reality, it is usually much more gradual.

A child may first learn simply to tolerate a new food on the table.

Next they might touch it.

Then smell it.

Then lick it.

Eventually they may hold it in their mouth before spitting it out.

Only much later do they swallow it.

Each of these tiny steps represents progress.

Therapists often describe this as systematic desensitisation.

The child slowly learns that new foods are safe.

Repeated exposure gradually reduces anxiety.

The process resembles treatment for phobias.

Nobody expects someone with a fear of spiders to begin by holding a tarantula.

Instead, they gradually become comfortable with increasingly challenging situations.

Eating works in much the same way.

 

Why repeated exposure changes the brain

Parents often become discouraged after offering a new food ten or twenty times without success.

Unfortunately, that may not be nearly enough.

Research on food acceptance suggests that some children need dozens—or even hundreds—of exposures before a new food becomes familiar.

Every successful exposure teaches the brain something important:

"Nothing bad happened."

Over time, anxiety decreases.

Novelty decreases.

The food becomes part of the child's "safe" repertoire.

This is why consistency matters so much.

Small gains repeated hundreds of times eventually become major changes.

The six-year success story that opened this article probably consisted of thousands of tiny victories that, on their own, hardly seemed worth celebrating.

Together, they transformed a life.

 

Case histories teach us what clinical trials cannot

Clinical trials tell us what usually happens.

Individual families remind us what is possible.

The parent whose story inspired this article did not achieve success in six weeks.

They achieved it in six years.

That distinction matters.

Modern medicine often expects rapid results.

Parents understandably hope that one supplement, one therapy or one new technique will produce dramatic improvements within a few months.

Development rarely works that way.

Children learn through repetition.

Brains change through repetition.

Skills improve through repetition.

Eating is no different.

Some children will improve quickly.

Others will take years.

The important thing is that progress remains possible.

 

Progress Is Measured in Years

One reason families abandon feeding programmes is that they judge progress too soon.

Imagine expecting a child to learn the piano after six lessons.

Or expecting fluent reading after one month at school.

We would never make those assumptions.

Yet many people expect eating habits to change within weeks.

Instead, it is more realistic to ask:

"Is my child eating more different foods this year than last year?"

That question shifts the focus away from daily frustrations and towards long-term development.

The family who achieved success over six years almost certainly experienced long periods where nothing appeared to change.

But change was happening.

It was simply happening slowly.

 

Conclusion

Reading the six-year success story and then reading the Stanford trial left me with the same conclusion.

Parents matter.

Not because they caused ARFID.

Not because they are expected to fix it overnight.

But because they are uniquely placed to help their child improve every single day.

An ARFID diagnosis should never be interpreted as a prediction of lifelong eating difficulties.

Instead, it should be viewed as the starting point for understanding why eating has become difficult and for developing a structured plan to improve it.

For some children, that means treating reflux, constipation or nutritional deficiencies.

For others, it means addressing anxiety or oral-motor problems.

For almost all children, it means creating repeated opportunities to experience new foods without fear.

The therapist may design the programme.

The doctor may identify underlying medical problems.

But it is parents who provide the thousands of moments in which change actually happens.

As the family who inspired this article discovered, those moments accumulate.

One new food becomes two.

Two become ten.

Ten become a varied and healthy diet.

It may take years.

There will almost certainly be setbacks.

Progress may be frustratingly slow.

An eight-year-old with a poor diet and sloppy handwriting can become a teenager with a healthy diet and neat handwriting. Much depends on empowering parents with the knowledge, confidence and practical strategies to guide that journey.

As you embark on that journey, choose your community wisely. Social media can be an invaluable source of shared experience, but it can also become an echo chamber of low expectations. Look for communities that acknowledge today's challenges while continuing to believe in tomorrow's possibilities. Surround yourself with people who encourage evidence-based action, persistence and hope, rather than resignation.

Today's restricted eater does not have to remain tomorrow's restricted eater.