Summary

  • Kent Kiehl’s brain-scan research and courtroom testimony operate within a self-reinforcing institutional loop in which citation volume generates funding, funding enables more research, and courtroom deployment signals legitimacy without requiring individual-level predictive validation.
  • The same MRI and genetic evidence presented to mitigate Amos Joseph Wells’s culpability was repurposed by the prosecution to argue future dangerousness, illustrating that brain data functions as a narrative vessel rather than an interpretive constraint.
  • Critics including MIT neuroscientist Satrajit Ghosh, Columbia psychiatrist Paul S. Appelbaum, and sociologist Oliver Rollins identify the ecological fallacy, the absence of base-rate analysis, and confounded sampling from racially skewed prison populations as structural deficiencies in the research’s courtroom application.
  • Funding continuity through a retracted 2018 paper and sustained NIH support exceeding $41 million since 2005 indicate the institutional loop lacks a countervailing feedback mechanism that would discipline the system when scientific claims fail.

Amos Joseph Wells, a Black man on Texas death row since 2016, was sentenced to die after his court-appointed defense attorneys hired Kent Kiehl’s company Mindset to coordinate scientific experts and presented MRI scans and genetic testing arguing Wells was biologically predisposed to violence. Defense counsel told jurors: “Amos didn’t ask for his genetics, he didn’t ask for the brain he got.” The prosecution took the same evidence and argued future dangerousness: “One thing they all agreed on is this guy’s dangerous. He’s never going to not be dangerous.” The jury sentenced Wells to death. Cooley LLP has petitioned the U.S. Supreme Court, arguing the trial was tainted by pseudoscience and racial bias; the NAACP and 29 scientists and lawyers have filed supporting briefs. The case crystallizes a broader pattern: brain-based evidence has become a structural feature of serious felony trials across the United States, and the institutional dynamics sustaining it operate largely independently of the evidence’s predictive validity.

The reinforcing loop

The trajectory of brain-based evidence in U.S. courts exhibits a reinforcing dynamic. Kiehl’s mobile MRI scans of incarcerated populations feed what his university webpage describes as “the world’s largest database of brain data,” which generates publications; those publications enter courtrooms, accumulating citations across thousands of cases; the citation volume signals legitimacy to judges, funders, and the defense bar; that legitimacy attracts more funding — Kiehl has received more than $41 million in NIH funding since 2005, including over $3 million in grants between 2025 and 2026 — which enables more research and more courtroom testimony through Mindset, which has been involved in more than 200 capital cases. Each circuit amplifies the signal without requiring the output to be independently validated at the level where it is actually deployed: the individual defendant. Systems theorists, following Donella Meadows, term this a reinforcing archetype.

Several secondary mechanisms strengthen the loop. Kiehl’s judicial education work — lecturing to “probably every federal judge in America” through the MacArthur Foundation’s Research Network on Law and Neuroscience — normalizes the evidence within the professional culture of the bench. Mindset, co-founded with his attorney wife in 2014, creates a commercial infrastructure that converts research into legal services. The defense bar’s need for novel mitigation strategies in capital cases — a need sharpened after Atkins v. Virginia (2002), which barred the execution of intellectually disabled persons and shifted the intellectual-disability framework — provides sustained demand. Brain scans offer defense attorneys a visually compelling, authority-laden counter-narrative to the prosecution’s case regardless of whether the underlying science can bear the weight placed on it.

A critical structural feature is the absence of a countervailing feedback mechanism. A retracted 2018 paper in Psychological Medicine — which Kiehl attributed to a student’s data-handling errors, a characterization the Cambridge Core retraction notice documents — produced no observable reduction in funding or courtroom deployment. Continued NIH support despite the retraction suggests the funding mechanism is insensitive to correction signals that in a self-correcting system would tighten the loop’s parameters.

Prevalence and institutional reach

Brain-based evidence appeared in more than 2,800 published judicial opinions between 2005 and 2015, according to a 2019 study reported by the Guardian. The study estimated neurological arguments for reduced criminal responsibility appeared in roughly 10–12% of U.S. murder trials and about 25% of death-penalty trials. Across all serious felony cases, 40% referred to brain-based evidence, a figure the study’s authors said “likely underrepresent[s]” actual prevalence. These specific figures could not be independently reproduced through available verification sources and should be treated as sourced to the Guardian’s reporting rather than externally confirmed.

Competing explanations for the adoption pattern

Four hypotheses attempt to explain why brain-based evidence has penetrated the courts despite scientific controversy. Each maps onto a different link in the reinforcing loop and implies different intervention points. The most parsimonious account combines the demand-side driver with the sampling-bias driver; funding continuity sustains the loop regardless of which combination holds, because the countervailing feedback that would normally discipline the system is absent.

Institutional demand (strongest-supported): The defense bar needs mitigation evidence courts will accept and juries will consider; brain scans fill that need. The 40% prevalence figure in serious felony cases is consistent with broad defense-attorney adoption. Mindset’s involvement in more than 200 capital cases indicates a functioning market. This hypothesis operates on the loop’s courtroom-credibility link; if correct, interventions should target defense-attorney incentives and courtroom admissibility standards.

Scientific validity: The science is sufficiently valid for courtroom use. Partially supported by Kiehl’s characterization of peer-reviewed, publicly funded research. Weakened by the retracted paper, by Ghosh’s comparison to phrenology — “They were measuring bumps and saying, ‘We think this person has this behavior because of these bumps’” — by Appelbaum’s observation that the deterministic approach “never works out,” and by the absence of published outcome data showing brain evidence produces systematically different sentencing outcomes. This hypothesis addresses the loop’s research-to-publication link.

Political and institutional capture: The field persists because it aligns with cultural narratives about biological determinism and because institutional infrastructure — NIH funding, judicial education, a commercial firm — has enough mass to sustain itself. Partially supported by funding continuity through the retraction, precisely the event that should have disrupted the funding link in a self-correcting loop. Would require additional evidence, such as internal communications showing political motivation, to rank above the demand hypothesis. Targets the funding-to-research link and the absent countervailing feedback.

Confounded sampling (Rollins): Studying incarcerated people’s brains cannot distinguish neurobiological traits from the effects of incarceration; the prison population’s racial and socioeconomic composition means the data reflects who gets locked up rather than who commits crimes. Sociologist Oliver Rollins identified that neuroscience has no way to account for this. This confound undermines the loop’s foundational research link, because if the observed brain differences are products of incarceration rather than predispositions toward crime, the entire evidentiary chain from scan to courtroom testimony rests on a causal inversion.

Critical missing evidence: Among the more than 200 capital cases where Mindset was involved, how often did brain evidence contribute to a sentence other than death? Without this data, neither the claim that the evidence helps defendants nor the claim that it backfires can be substantiated.

The racial disparity problem

Rollins identified that the prison population is shaped by systemic racism, so any brain study conducted in prisons will disproportionately study people of color. The article reports that Black people are incarcerated at more than five times the rate of white people in Ohio, and that racial disparities have amplified since the 1970s — figures confirmed by multiple independent sources including a Health Policy Ohio analysis, the Vera Institute, and the Prison Policy Initiative. Kiehl acknowledged that the criminal justice system is biased but characterized the bias as directed toward “people of low socioeconomic status, not race.” Rollins’s critique and the reported data suggest the socioeconomic explanation does not fully resolve the racial disparity; the Ohio figures are consistent with a structural racial disparity that socioeconomic variables alone do not account for. This dynamic maps onto what systems analysts describe as a “shifting the burden” archetype: the criminal justice system’s racial selection bias feeds the brain databases, whose outputs circulate as seemingly neutral biological explanations for violence, which in turn can reinforce the sentencing patterns that produce the biased sample.

As Matthew Kutcher, a partner at Cooley representing Wells, put it: “If you tell a jury that a Black man is genetically wired for violence, you’re inviting them to see him through racist stereotypes, not as an individual.”

Coherence failures in the courtroom deployment

Bidirectional vulnerability. The same evidence that was introduced to mitigate Wells’s culpability was used by the state to justify his execution. This is not a contradiction at the level of science but at the level of what the science is asked to do. If brain data are interpreted as a static, deterministic marker, both the defense’s mitigation theory and the state’s aggravation theory follow from the same data, and the scan does not resolve which inference to draw. The evidence functions as a vessel for narrative rather than a constraint on interpretation. The Brian Dugan case — where Kiehl described Dugan as “just utterly and completely psychopathic. A perfect – I mean, I hate to say this word – specimen” yet the jury still returned a death sentence — reinforces this pattern.

Ecological fallacy at the individual level. Even if brain-structure differences correlate with psychopathic traits or violent behavior at the group level — a claim critics challenge — applying that correlation to a specific defendant requires acceptable individual-level diagnostic accuracy. Population-level neuroscience identifies associations; courtroom application requires individual-level diagnostic accuracy. The bridge between the two does not exist in the published literature as described.

Process-for-predictive validity substitution. Kiehl has defended his work as “peer-reviewed, publicly funded, and goes through a rigorous ethical review process.” Each is a property of the research process; none addresses whether the research produces reliable predictions about individual behavior. A methodologically rigorous study of a group-level phenomenon does not become an individually diagnostic tool by virtue of having passed ethical review.

Presumption fallacy. Appelbaum’s critique — that the deterministic approach “never works out” because it seeks “a single simple explanation for complex, multi-determined behavior” — identifies a presumption that criminal behavior can be reduced to brain structure and genetics. This presumption is introduced through the framing itself: once the defendant is presented as someone whose brain predisposes him to violence, the social and institutional context recedes from view.

The base-rate gap

The most consequential analytical absence in the courtroom discourse is the base-rate problem. Individual-level diagnostic use of brain-based evidence requires knowing the base rate of the relevant trait in the population and the false-positive and false-negative rates of the assessment. A worked illustration: a brain-based measure claiming 90% sensitivity and 90% specificity — figures well beyond what any brain-scan study has demonstrated for violent behavior — applied to a population where the relevant trait has a 1% base rate yields a positive predictive value of roughly 8%. In a group of 1,000, that produces 9 true positives and 99 false positives out of 108 positive results, meaning approximately a 92% chance that a positive result is wrong. The actual accuracy of brain-based behavioral prediction falls far below these hypothetical thresholds. Ghosh and Appelbaum both gesture at this gap without naming its formal structure.

Leverage points and the absent gatekeeper

The Wells case illustrates the absence of an institutional gatekeeper that would constrain the interpretation of brain data to one inferential direction. Three interventions target different links in the reinforcing chain:

Stricter Daubert standards — requiring demonstration of individual-level predictive validity — would reduce the courtroom signal that feeds back into the funding and credibility loop. This represents the highest-leverage intervention and requires no new legislation or new science.

Research design reform — adequate non-incarcerated representation, distinguishing neurobiological traits from incarceration effects, accounting for racial and socioeconomic composition — would weaken the research foundation without eliminating it.

Alternative mitigation strategies in the defense bar that do not carry the bidirectional risk the Wells case illustrates would reduce the demand that feeds the loop.

The NAACP and 29 scientists and lawyers filing briefs in Wells’s Supreme Court case marks the presence of a balancing loop that could, if it leads to a gatekeeping ruling, restrict admissibility and limit the reinforcing loop’s reach. The system has not yet resolved whether its balancing mechanisms are strong enough to counteract the institutional and economic momentum behind the scan.

Verification notes

The 2018 retraction of a Kiehl-co-authored paper in Psychological Medicine was confirmed through the Cambridge Core retraction notice. Ohio’s Black-to-white incarceration rate exceeding 5:1 was confirmed through Health Policy Ohio analysis, the Vera Institute, and the Prison Policy Initiative. The 2019 study’s specific figures (2,800 judicial opinions, 10–12% of murder trials, 25% of death-penalty trials) and the $41 million NIH funding figure could not be independently reproduced and should be treated as sourced to the Guardian’s reporting rather than externally confirmed.

Analytical techniques used in this piece

This analysis applies the methods below. Each links to a short, plain-English explainer you can read and reuse.

Coherence Audit
Tests whether an argument hangs together — spotting contradictions, gaps, and circular reasoning.
Analysis of Competing Hypotheses
Scores rival explanations by how well each fits the evidence, weighting the diagnostic items (Heuer).
Systems Dynamics (Causal)
Models the feedback loops and delays that drive a behavior over time.
Confirmation Bias
Seeking and overweighting the evidence that confirms what one already believes.