The Causes & Treatments for Autism | Dr. Karen Parker
- Guest
- Dr. Karen Parker
- Read time
- ~4 min
- Uploaded
- 11 Dec 2023
- Duration
- 2:56:28
- Views
- 483.3K
KEYPOINTS
- 01
Autism prevalence is now about 1 in 36 U.S. children
Parker, a Stanford researcher, says the rate rose from 1 in 44 just a couple of years earlier. She calls the increase multifactorial: more awareness and routine early screening explain part of it. Autism is male-biased, and it is highly heterogeneous: 'If you've met one kid with autism, you've met one kid with autism.'
- 02
Diagnosis rests on two core features
Under the DSM-5, an expert (usually a psychiatrist or psychologist) looks for pervasive social interaction challenges and restricted, repetitive behaviors. Many autistic people also have anxiety, sensory challenges, seizures or sleep disorders. For most people, autism has lifelong impact, though some outgrow the diagnosis and some respond well to behavioral therapy while others see minimal benefit.
- 03
Genes matter a lot, and traits run on a continuum
Parker estimates that about 40–80% of autism is genetic, with roughly half linked to common genetic variants. Younger siblings of autistic children are at higher risk, which researchers use in prospective 'baby sib' studies. Autistic traits measured by the Social Responsiveness Scale or Autism Quotient are spread through the general population. They are higher among people in intense STEM fields, even without a diagnosis.
- 04
Environmental risk factors interact with biology
Associated risks include factors such as maternal illness during pregnancy. These come mainly from large epidemiological studies, since experiments in developing children are rarely possible. Huberman raises Pasko Rakic's mouse work linking repeated prenatal ultrasound to neuronal migration errors. Huberman emphasizes 'perhaps' and notes that the idea never gained traction.
- 05
Only two drugs are FDA-approved for autism, and both are antipsychotics
They treat associated features like irritability rather than core social symptoms and carry side effects like weight gain. That gap is why Parker focuses on neuropeptides that regulate social behavior: oxytocin and vasopressin, which are ancient, evolutionarily conserved molecules made in the hypothalamus.
- 06
Oxytocin may help a subgroup, not autism in general
In a study of about 200 children (autistic, unaffected siblings and controls), lower blood oxytocin tracked greater social difficulty in all groups, but it did not separate autism from non-autism. In Parker's double-blind trial with child psychiatrist Antonio Hardan, children given oxytocin twice daily for four weeks benefited if they started with low oxytocin. A later multi-site phase 3 trial found no overall benefit. Parker thinks measurement problems may have obscured a responsive subgroup, and she notes that pediatric oxytocin has looked relatively safe.
- 07
Vasopressin looks like the stronger bet
Parker says that if she had to bet, she'd bet on vasopressin. In rhesus monkeys at the UC Davis primate center, naturally 'low social' monkeys, which she stresses don't have autism, could be classified almost perfectly by vasopressin in cerebrospinal fluid (CSF), though not in blood. Levels were stable within individuals over time, a requirement for a biomarker. In autistic children, CSF vasopressin tracked social symptom severity but not repetitive behavior.
- 08
Low CSF vasopressin appeared before symptoms
In banked newborn CSF samples, infants who were later diagnosed with autism already had low vasopressin, while oxytocin didn't differ. The samples came from John Constantino, whose team traced about 2,000 paper records. Parker emphasizes it was a small sample that needs replication. She speculates that many susceptibility genes might converge on a few pathways, vasopressin among them.
- 09
The vaccine–autism idea is not supported
Parker says the original study was debunked: its author lost his medical license over fabricated data and the paper was retracted. Every study she knows of finds no link between vaccination and autism. Most scientists and doctors she knows vaccinate their own children.
EXAMPLES
- 01
First vasopressin treatment trial in autistic children
In a randomized, double-blind, placebo-controlled trial (17 children on drug, 13 on placebo), intranasal vasopressin twice daily for four weeks improved social abilities on three measures: parent report, clinician evaluation and lab tests. A subset also showed less anxiety and fewer repetitive behaviors. Parker warns that dosing matters and adverse effects could be severe, so no one should try it outside a trial.
- 02
Voles and day length
Monogamous prairie voles bond and co-parent, unlike meadow voles, where mothers raise pups alone. Parker noticed that male meadow voles housed on winter-like day lengths began hanging out with females and helping with litters. That observation drew her into vasopressin and social behavior.
- 03
Probiotics in mouse models
In genetically modified mice with social impairments and low oxytocin, giving a probiotic normalized social behavior and raised hypothalamic oxytocin, and in one study vasopressin too. Cutting the vagus nerve blocked the rescue, which shows the effect runs from gut to brain through the nerve. These are animal findings.
- 04
Why primate models matter
Thalidomide, a morning-sickness drug, had its toxicity testing done only in mice. Children were then born with severe limb abnormalities. When it was later tested in primates such as marmosets, the defects appeared, and Parker notes that even a single primate test might have prevented the harm. MPTP, the synthetic heroin that caused overnight parkinsonism in people, also had little effect in mice. Monkey social behavior, including affiliative gaze and grooming, maps onto human behavior in ways mouse behavior doesn't.
OTHER RELEVANT
- 01
Early behavioral intervention still matters
Diagnosis can wait 12–18 months at clinics during critical early windows. That is why Parker wants faster lab-based screening tests, such as blood or CSF biomarkers. Evaluation should come from qualified experts, not self-diagnosis.
- 02
Neuropeptide treatments need a prescription and trial evidence
Oxytocin spray requires a physician's prescription and isn't approved for autism. Vasopressin results come from a small trial that is being extended to larger groups. Parker suggests stratifying patients by baseline levels.
- 03
Funding delays have costs
Huberman notes that three to five years of grant writing can stand between a sound hypothesis and a trial, while children age through critical developmental windows.