2026 · JAMA Psychiatry
Changes in Genetic Contributions to ASD and ADHD by Year of Diagnosis
Verdict
UNSCIENTIFIC
Neither the paradigm nor the study inside it passes. The verdict is unscientific.
1·Study methodology · inside the paradigm
UNSCIENTIFIC
Does not follow the scientific method within Genomics / DNA Model.
2·Paradigm · Genomics / DNA Model
UNSCIENTIFIC
Fails the tests. This decides the verdict.
How the verdict is decided
The verdict grades the whole picture: the paradigm this study assumes, and how the study was carried out inside it. A study can follow the scientific method rigorously and still be unscientific, because rigor inside a paradigm only shows the conclusion was reached carefully. It cannot verify the premise the paradigm rests on. So the paradigm decides the verdict. The study’s own score is kept because it shows how the conclusion was built.
1·Study methodology
Did this study test its claim with methods that are independent, falsifiable, and non-circular?
UNSCIENTIFIC
This paper claims that polygenic scores, computed from genotyping arrays and statistical models, show declining genetic risk among people diagnosed with ASD or ADHD over time, supporting the idea that broadened diagnostic criteria explain rising diagnosis rates. All three tests fail.
Independently Verifiable
Fail
Genotyping arrays produced fluorescent signals and software interpreted those signals as genetic variants using the DNA model, so the genetic risk profile only exists as the model's reading of a signal.
Falsifiable
Fail
If the polygenic scores had shown no change across years, the study would have concluded that genetic risk was stable and environmental factors drove the increase, leaving the genomics framework untouched either way.
Non-Circular
Fail
The polygenic scores were built from GWAS results that assume the DNA model, the genotyping probes were designed from an assembled reference genome, and the psychiatric diagnoses used to define the cohorts are themselves model-dependent categories.
Why
They took people diagnosed with ASD or ADHD between 1994 and 2016 and sorted them by year of diagnosis. Genotyping arrays produced fluorescent signals from their biological samples. Software interpreted those signals as genetic variants using the DNA model. A computer combined thousands of model-identified variants into a single number called a polygenic score for each person. They compared the average score across diagnosis years and found it declined over time. They called this declining genetic risk. The entire measurement is the model narrating its own output. A person who rejects the DNA model sees fluorescent signals and a computed number. They do not see genetic risk for autism. The study never tested whether polygenic scores measure anything real. It assumed they do and processed the assumption through regression models and simulations. If the scores had not changed, the field would have called it stable genetic architecture and pointed to environmental factors. The genomics framework survives every possible result. The diagnostic categories that defined the cohorts are themselves psychiatric model constructs. Every layer of this study presupposes the model it claims to inform.
2·Paradigm · Genomics / DNA Model
Does the framework this study assumes pass the three tests?
UNSCIENTIFIC
Independently Verifiable
Fail
The genomics paradigm claims that fluorescent signals from genotyping arrays represent a hereditary code that determines psychiatric risk, but someone who rejects that model sees only signals on a screen.
Falsifiable
Fail
When polygenic scores fail to predict outcomes, the field calls it missing heritability or gene-environment interaction, so no result can disprove the claim that DNA variants cause psychiatric conditions.
Non-Circular
Fail
The paradigm defines psychiatric conditions using diagnostic categories, searches for genetic variants associated with those categories using probes built from an assembled genome, then uses the variants to compute risk scores for the same categories.
Why
The genomics paradigm claims that DNA contains a hereditary code that determines who develops psychiatric conditions. The instruments that produce the data are genotyping arrays and sequencers. They emit fluorescent signals. The model converts those signals into variants, genes, and polygenic scores. Without the model, the output is lines on a screen. The paradigm has never been independently verified because every observation it produces requires the model to run first. When predictions fail, the field does not question whether DNA variants cause psychiatric conditions. It calls the gap missing heritability, invokes gene-environment interaction, or says more variants need to be discovered. The paradigm cannot die because no result is allowed to count against it. The reference genome was assembled using the model. The probes were designed from that assembly. The GWAS studies that generate polygenic scores search for associations between model-defined variants and model-defined diagnoses. The tool was built to find what the model already named. This paper is a perfect example. It computed risk scores from model-dependent instruments, applied them to model-defined diagnostic categories, and reported the output as evidence about genetic risk. The premise went in untested and came out looking like a finding.
From the paper
Polygenic scores for psychiatric (ADHD, ASD, depression, bipolar, and schizophrenia) and cognitive-behavioral (addiction, educational attainment, IQ, neuroticism, and risk-taking) outcomes were used to capture genetic risk profiles of diagnosed individuals.
A more recent ADHD diagnosis was associated with decreased genetic risk for ADHD as shown by difference in mean polygenic score (β estimate per 10-year increase, −0.06 SDs; 95% CI, −0.09 to −0.03 SDs; P = .001)
Findings support broadening diagnostic criteria as an explanation for increasing rates, with implications for understanding changes in risk factors and clinical practice.
Source
Changes in Genetic Contributions to ASD and ADHD by Year of Diagnosis
LaBianca, Sonja; Lousdal, Mette Lise; Dybdahl Krebs, Morten; Sahlholdt Hansen, Ole; Georgii Hellberg, Kajsa-Lotta; Lundberg, Mischa; Østerby Sørensen, Johanne; Gådin, Jesper R.; Ohlsson, Henrik; Grove, Jakob; Børglum, Anders; Agerbo, Esben; Werge, Thomas; Albiñana, Clara; Vilhjálmsson, Bjarni J.; Kendler, Kenneth S.; Plana-Ripoll, Oleguer; Schork, Andrew J.; iPSYCH Study Consortium and Autism Spectrum Disorder Working Group of the Psychiatric Genomics Consortium; Nordentoft, Merete; Mors, Ole Mors; Mortensen, Preben Bo; Demontis, Ditte; Grove, Jakob; Starnawska, Anna; Damm Als, Thomas; Buil, Alfonso; Rosengren, Anders; Ingasson, Andres; Helenius, Dorte; Zetterberg, Richard; Bøcker Pedersen, Carsten; Christensen, Jakob; Pedersen, Liselotte; Giørtz Pedersen, Marianne; Bybjerg-Grauholm, Jonas; Bækvad-Hansen, Marie; Havdahl, Alexandra; Carracedo, Angel; Breetvelt, Elemi J.; Corfield, Elizabeth C.; Hakonarson, Hakon; Dziobek, Isabel; Underwood, Jack; González-Peñas, Javier; Duan, Jinjie; Weiss, Lauren; Tebartz van Elst, Ludger; Gandal, Michael J.; Andreassen, Ole A.; Anney, Richard J.L.; Scherer, Stephen W.; Kuo, Susan S.; Thomas, Taylor R.; Pillalamarri, Vamsee; Ahlqvist, Viktor H.
2026 · JAMA Psychiatry
10.1001/jamapsychiatry.2026.1450
Scored from abstract only · Rubric 1.0