Does Early Methamphetamine Abstinence Change Brain Structure Without Detectable Myelin Loss? Evidence From 27 Adults

Short answer: A 2026 observational study of 27 adults found that 13 people in early remission from methamphetamine use disorder performed less efficiently on some Tower of London planning measures and had thinner or smaller cortical regions on structural MRI than 14 comparison participants. The same study did not detect significant group differences in its MRI-derived myelin estimates, subcortical volumes, or several other planning outcomes. The study did not administer methamphetamine, measure a contemporaneous parent-drug concentration in legally collected venous whole blood, or test actual driving or courtroom behavior. These are group-level associations, not an individual intoxication test or proof that methamphetamine alone caused a particular person’s brain finding.

The underlying original empirical report is Bristow and colleagues’ Multimodal MRI Reveals Brain Structural Differences and Executive Dysfunction in Early Methamphetamine Abstinence, published in Brain and Behavior in 2026: the complete original study (DOI 10.1002/brb3.71288). Its informative contrast is not simply an abnormal brain scan versus a normal one. It is the coexistence of selected cortical and planning differences with numerous genuinely null or unresolved measurements in a small, nonrandomized sample.

What is the narrow forensic question answered by this study?

The appropriate question is whether a particular recruitment group—adults diagnosed with methamphetamine use disorder and reporting less than a month of abstinence—differed from locally recruited controls in cortical MRI anatomy, quantitative myelin-sensitive MRI, and a standardized laboratory planning task. That question is narrower than whether an identifiable person used methamphetamine on a particular date, was under its influence at a specified moment, or suffered permanent brain damage from a known dose. The investigators performed an observational case–control comparison. The exposure of interest was an existing clinical and self-reported history, not a controlled intravenous, intranasal, oral, or smoked drug administration during the protocol.

For a forensic reader, this distinction is foundational. Structural MRI quantifies anatomical features using image processing. A computerized planning task measures performance under one set of instructions. Neither technology chemically identifies methamphetamine in an individual’s blood specimen or supplies a validated cut-off for impaired driving, truthfulness, capacity, or responsibility.

How many independent participants were studied?

The sample consisted of 27 independent people: 13 with methamphetamine use disorder who were in early abstinence and 14 comparison participants without a methamphetamine-use history. These were two recruited groups, not a randomized drug trial. The methods identified current intoxication among the exclusion criteria. The authors described MRI at 3 tesla and one Tower of London cognitive assessment per participant. Cortical vertices, brain regions, task trials, estimated myelin values, and model outputs are repeated or derived observations within those people; they cannot properly be added to the denominator as new biological subjects.

The groups were similar in reported mean age, but they were not perfectly matched in all potentially important characteristics. In the original participant table, mean education was 12.00 years in the methamphetamine group and 14.71 in controls, and reported prior head injury was present in five of 13 cases but no controls. The sample included relatively high rates of smoking and polysubstance histories among cases. Age and sex were included as imaging covariates, but that does not by itself eliminate confounding from education, head trauma, tobacco, sleep, other substances, childhood experiences, or selection into treatment.

What was actually done to participants?

Participants underwent high-resolution T1-weighted structural MRI and a separate myelin-sensitive MRI acquisition, using a 3.0-tesla system. FreeSurfer software segmented cortex and subcortical structures and estimated regional cortical thickness and volumes. A MAGiC acquisition processed with SyMRI provided model-based myelin-sensitive values, which were then assigned to white-matter regions. Cortical vertex comparisons applied age and sex covariates with cluster-wise correction. Some subsequent regional models examined adjusted associations between self-reported years of use and estimated cortical volumes.

The Tower of London task required planning a sequence of ball movements to achieve a target configuration in a limited number of moves. Its outputs included first-attempt accuracy, total task score, solution time, execution time, and attempts. Those outcomes do not all measure the same mental operation. Planning efficiency, time spent forming a solution, time spent executing it, and accuracy on first attempt can move in different directions.

No controlled methamphetamine dose or route was administered during the experiment. No blood draw was reported for chemical confirmation or mass-spectrometric quantification of methamphetamine, amphetamine, or other analytes at the MRI or cognitive-testing visit. The magnet, image-analysis software, and task computer were the principal instruments relevant to the reported findings; none is a forensic toxicology mass spectrometer.

What did the Tower of London experiment actually show?

On three reported indices, the group with methamphetamine use disorder had less efficient task performance. The execution component of correct solutions took longer, the total number of attempts was higher, and fewer solutions were correct on the first attempt. These differences were reported with unadjusted nominal p values of .01, .02, and .01, respectively. The raw comparison was not a uniform deficit on every test parameter.

The combined task total score was not statistically different at the source’s significance threshold (p=.18). Neither mean solution time on correct solutions (p=.69), total first-move time (p=.36), mean first-move time (p=.26), nor the calculated impulsivity index (p=.27) showed a significant group difference. These null results matter: selecting only the three significant rows would give readers an inaccurate impression of a globally failed executive-function test.

At the same time, a nonsignificant group difference does not demonstrate that the groups are equivalent. With 13 and 14 people, wide distributions and limited power can conceal small or even meaningful effects. The original source measured this version of a planning task, not the complete range of memory, attention, executive control, intelligence, interpersonal reliability, or everyday occupational performance.

Numerical evidence table 1: planning outcomes versus actual null comparisons

Source-faithful interpretation: values are original group means ± standard deviations from the published Tower of London table. These are behavioral scores and time measurements, not drug concentrations. A p value compares a particular statistical model or contrast; it does not measure the probability that any individual was intoxicated, dishonest, or incapable of planning.

Where did structural MRI show differences?

Age- and sex-adjusted cortical analysis identified two regions of reduced thickness among cases relative to controls: left lateral occipital cortex and right lingual gyrus. The original cluster-corrected p values were .015 and .002. These areas participate in aspects of visual processing, but the imaging result does not prove a measured visual deficit in each subject. It reports differences in MRI-derived anatomy across two small groups.

For cortical volume, the reported clusters were right superior frontal cortex and right lingual gyrus. The source reports a cluster-corrected value below .01 for the superior frontal region and .018 for the lingual region. Cortical thickness and volume are related anatomical quantities but not interchangeable: a thickness estimate, an area, and an intracranial-volume-adjusted regional volume carry different units and may respond differently to segmentation errors, head size, and sampling.

The researchers also explored correlations between self-reported years of methamphetamine use and regional cortical volumes. They described a negative relation for the right caudal middle frontal region and several additional areas. This was observational exposure history, not chemically verified lifetime methamphetamine dose. Years of use should not be labeled an administered dose, a pharmacokinetic area under the curve, or a contemporaneous blood result.

Did the study establish that methamphetamine caused the cortical differences?

No. The imaging results are compatible with a group-level association, not a randomized causal test. The study did not have pre-use brain scans for the same people, a controlled methamphetamine exposure and longitudinal baseline, or comprehensive resolution of education, head injury, tobacco, polysubstance exposure, mental-health history, sleep, and other variables. These factors could influence both selection into the methamphetamine-use group and the anatomical or task measurements.

Even when an MRI cluster survives a particular correction procedure, causation is a separate question. The finding cannot identify which exposure produced it, when it arose, whether it preceded the drug history, or whether it will reverse with longer abstinence. The paper’s discussion proposes biologically plausible mechanisms; those mechanisms were not established by measuring dopamine, synaptic inflammation, neuronal death, or methamphetamine molecules in the participants’ cortex.

What did the myelin-sensitive MRI results fail to detect?

The study’s myelin-sensitive sequence did not identify significant group differences across the evaluated white-matter tracts. The discussion likewise describes no significant cortical or global myelin-volume differences and no significant correlations between tract values and reported years of methamphetamine use. A descriptive global mean was 166.73 in the methamphetamine group versus 172.10 in controls, with p=.81. This was an MRI-derived myelin-related estimate, not a microscopic count of myelin sheaths or an assay of a myelin-specific protein.

The word preserved can be tempting, but a negative test in a small sample cannot conclusively establish intact myelin in every participant. A synthetic-MRI signal may also reflect modeling assumptions and sensitivity limits. It is more accurate to say that this protocol did not detect an above-threshold group difference in those estimates. Similarly, the investigators reported no significant subcortical structure differences; that is a study-specific null finding, not a claim that striatum or hippocampus is never affected by methamphetamine.

Numerical evidence table 2: cortical positives beside measured MRI negatives

The original paper reports MRI-derived measurements and cluster-level inference; the cluster area is not a brain-region volume difference, and the quoted myelin index does not have units of blood methamphetamine. The source presents additional anatomical regions, myelin tracts, and exposure-duration correlations. This deliberately selective table contrasts source-confirmed positive and true null findings without treating anatomical image features as independent human observations.

What exactly was chemically measured, and in which matrix?

For the outcomes discussed here, no parent methamphetamine or amphetamine concentration was chemically quantified in a legal human venous whole-blood specimen. The case history defined participant grouping; the imaging pipeline generated estimates of cortical structure and myelin-sensitive parameters, while the Tower of London generated performance variables. These are not instruments or matrices for determining nanograms per milliliter of a drug.

Legally collected venous whole blood contains cells and plasma. Separated plasma and serum are different physical matrices and cannot be silently substituted for whole blood; an observed level in one does not acquire a validated conversion into the other merely because the drug name matches. Urine testing generally answers a different detection question; oral fluid, sweat patches, hair, breath, vitreous humor, breast milk, animal tissues, and artificial laboratory matrices each have their own sampling, contamination, and interpretive limitations. None of those served as a chemical methamphetamine concentration endpoint in this MRI study.

The fact that a participant reported abstinence is not a quantitative pharmacokinetic observation. Even a future verified toxicological result would still require a collection date, specimen identification, chain of custody, validated analyte-specific method, laboratory uncertainty, and clinical context before any case-specific inference. For specimen fundamentals, see the site’s already published guide to forensic biological matrices and its distinct guide to confirmatory drug testing.

Was this a drug-administration, blood-level, or pharmacokinetic study?

It was none of those. The researchers did not administer methamphetamine, d-amphetamine, racemic amphetamine, lisdexamfetamine, methylphenidate, or a placebo treatment. Therefore there was no investigator-assigned milligram dose, intravenous rate, oral absorption curve, timed serum concentration, half-life, clearance, or blood/plasma distribution ratio to report. It is scientifically incorrect to use the history of a participant who reported many years of methamphetamine use as if it were a measured dose or drug concentration.

It is equally incorrect to interpret cortical thickness in millimeters, a volume estimate, an image segmentation score, a cluster-corrected p value, or a timed puzzle response as a parent-drug analyte result. Neuroanatomical imaging can identify group contrasts; forensic analytical chemistry identifies and sometimes quantifies a specified molecule in a collected physical specimen. These two lines of evidence should complement rather than impersonate one another.

How should the genuine null results influence interpretation?

The null Tower of London total score and mean solution time indicate that the observed planning profile was selective, not a consistent decline on every endpoint. The absence of a detectable myelin difference under this specific protocol constrains a broad claim that every neural compartment must show measurable group damage. The subcortical null similarly discourages an unqualified statement that every reward or memory-related brain region was structurally different.

At the same time, a p value above .05 does not certify equality or normality. These were small groups with potentially substantial measurement variance, and multiple endpoints were examined. A fair interpretation states the outcome, the direction, its p value where provided, the correction procedure when relevant, and the limitations of the study’s design. An independent case expert should distinguish a true source-reported null from an endpoint that was never collected at all.

Why do sample size and baseline differences matter so much?

With only 13 cases and 14 controls, any one participant can meaningfully affect a mean, a regional regression, or an estimated association. The included participants were not a random sample of everyone who uses methamphetamine. They were people recruited through rehabilitation services in a particular New Zealand setting, with a particular demographic makeup and exclusion criteria. Generalization to a different age group, clinical history, jurisdiction, or biological specimen requires independent support.

Education differed significantly between the groups, and reported head injury occurred more frequently among cases. Tobacco and other drug histories were uneven as well. Those are not minor administrative details: they are alternate or contributing explanations worth testing before attributing a difference entirely to one drug. The authors acknowledged several unmeasured factors, including sleep, diet, childhood trauma, and broader health context. An adjusted group coefficient is not the same as having randomized these characteristics.

What does the study tell us about duration of methamphetamine use?

Among cases, the authors reported negative age- and sex-adjusted correlations between recalled years of drug use and selected regional cortical volumes. One cited right caudal middle frontal coefficient was r=−.754 with p=.001. Such an association asks whether larger reported durations tracked smaller MRI-estimated volumes within that sample. It does not establish a threshold, a dose–response law in chemically verified units, irreversible atrophy, or which person will show the finding.

Duration is an imprecise proxy for total exposure: it does not capture dose per episode, potency, changing use frequency, periods of abstinence, polydrug intake, route, or developmental factors. The published correlation table also contains internal inconsistencies between some printed confidence intervals and p values, an editorial caution that warrants careful source inspection. The safe scientific conclusion is that the reported exploratory correlations merit replication—not that years of use can reconstruct an individual’s measured drug level or legal state.

Does finding a smaller cortex mean a person was impaired at an incident?

No. A case-specific impairment proposition requires evidence tied to the individual’s relevant time, behavior, tasks, medications, exposure history, and a scientifically validated relationship to the legal question. This study had no actual on-road driving, simulated collision, field sobriety test, eyewitness identification task, courtroom testimony evaluation, or retrospective analysis of a specific injury or death. The Tower of London test was not validated in this study to distinguish intoxicated from unimpaired individuals or to identify whether a person was affected by methamphetamine at a particular hour.

Even when a group difference is genuine, distributions may overlap. A group’s average execution time does not prove how any individual member performed, much less how an entirely different litigant performed. Brain imaging is not a forensic identity assay for drug use. It also does not independently establish legal intoxication, diminished capacity, credibility, intent, or causation.

How should clinicians and forensic scientists use these findings?

For clinical research, the study suggests that early abstinence may be a useful period for investigating planning difficulties alongside frontal and visual-region MRI anatomy. It also makes myelin-sensitive measures valuable controls rather than guaranteed positives. Larger prospective studies with repeated within-person imaging, standardized head-injury and substance-use verification, broader cognitive batteries, and controls for education and health would be needed to determine persistence or recovery.

For forensic education, the study is a practical example of an endpoint boundary: MRI regional measurements, a task-based cognitive index, self-reported history, and quantitative drug toxicology address related but fundamentally different questions. An opinion should not cross those boundaries without validated bridges and facts from the actual case record. Objective chemical evidence, when available, must be interpreted under the specimen-specific conditions described in the laboratory documentation.

Seven scientific FAQs

1. Did the 2026 study prove methamphetamine shrinks everyone’s brain?

No. It compared groups and found differences in selected cortical clusters. It was not randomized, did not establish pre-use baseline scans, and was too small and confounded to prove individual or universal causation. Other regions and modalities did not show statistically significant differences.

2. How many humans were independently evaluated?

There were 27: 13 early-abstinent participants with diagnosed methamphetamine use disorder and 14 controls. Each participant’s many brain vertices, regional MRI estimates, and task responses are repeated or derived measurements, not additional independent participants.

3. Did the experiment measure blood methamphetamine or assign a dose?

No. It did not administer a drug or use a quantitative chemical assay to establish parent methamphetamine or amphetamine in legal human venous whole blood, separated plasma, serum, urine, oral fluid, sweat, hair, or another drug-testing matrix. Self-reported history cannot substitute for measured drug analyte and concentration.

4. Were all executive-function results abnormal?

No. Execution time, total attempts, and first-attempt accuracy differed by group, but total task score, solution time, first-move measures, and the calculated impulsivity index did not meet the study’s significance threshold. The study used one principal planning task, not a comprehensive forensic cognitive battery.

5. Did the study show myelin damage?

No significant group difference was detected by its particular synthetic-MRI-derived myelin measures. This negative finding does not demonstrate that every participant’s microscopic myelin is normal, nor does it prove the technique is equally sensitive to all possible forms of microstructural injury.

6. Can a structural MRI result reveal when someone last used methamphetamine?

No. The protocol supplied no validated individual timing model, pharmacokinetic sampling, or measured postdose analyte level. Cortical thickness, volume, and synthetic-MRI myelin values do not identify a last-use hour or a legal drug concentration.

7. Can these data prove that methamphetamine caused a crash, bad decision, or unreliable recollection?

No. Such events were not observed or experimentally evaluated. The group-level findings support cautious scientific discussion about planning and neuroanatomical association, not a personalized conclusion about impairment, causation, witness accuracy, or legal responsibility.

Source and final evidence boundary

This article is an original interpretation of Bristow, Tayebi, Condron, Emsden, Ngarimu, Kwon, McLanachan, Liu, Schierding, McHugh, Holdsworth, Mohi, Newburn, and Scadeng (2026), Multimodal MRI Reveals Brain Structural Differences and Executive Dysfunction in Early Methamphetamine Abstinence, Brain and Behavior 16(5):e71288. Read the primary open-access research article at DOI 10.1002/brb3.71288 for the full methods, tables, figures, and limitations. Published group differences deserve rigorous attention, but a source faithful interpretation also preserves its negative findings, recruitment confounders, analytical-matrix boundaries, and the absence of individual incident-level validation.