# Predictive single-dose test could guide ADHD treatment in adults > Source: > Published: 2026-08-13 16:00:54+00:00 Testing how a person responds to a single dose of stimulant medication might help doctors predict how well the treatment will work for them two months later. Researchers found that early changes in specific symptom scores accurately forecast long-term medication success in adults with attention-deficit/hyperactivity disorder. The study was published in the journal [Translational Psychiatry](https://doi.org/10.1038/s41398-025-03557-3). Attention-deficit/hyperactivity disorder, commonly known as ADHD, is a neurodevelopmental condition characterized by varying levels of inattention, hyperactivity, and impulsivity. Stimulant medications, such as methylphenidate, represent the primary pharmacological treatment for the condition. These drugs help regulate attention and behavior by enhancing the availability of dopamine and noradrenaline in the brain. This mechanism reduces background noise in neural circuits, enabling patients to better maintain focus and control impulsive behaviors. While stimulants reduce symptoms across large populations, individual patient responses vary widely. In standard clinical practice, doctors often prescribe these medications using a trial-and-error approach. Patients take a drug for weeks or months to see if it works and whether it causes intolerable side effects. If a medication fails, the patient must slowly taper off and try another option. This process can delay relief and prolong the academic, professional, and personal challenges patients face in their daily lives. To bypass this waiting period, researchers have been searching for ways to predict treatment success before a patient commits to a long-term prescription. Previous investigations hinted that measuring brain activity after just one dose of medication could indicate its long-term effectiveness. A research team led by Valeria Parlatini, a clinical academic at the University of Southampton and King’s College London, wanted to know if this predictive capability also applied to simpler, more accessible clinical tests. They designed a study to see if standard symptom questionnaires and computer-based behavioral assessments could forecast longer-term treatment outcomes. The research team recruited 60 adult males diagnosed with ADHD. At the beginning of the study, each participant attended two testing sessions. In one session, they received a single 20-milligram dose of methylphenidate. In the other session, they received a placebo capsule. Participants did not know which pill they were taking on which day, ensuring their expectations did not influence their performance. During both sessions, the researchers evaluated the participants using two tools about an hour after they took the pill, allowing the medication time to enter their bloodstream. The first tool was a self-report questionnaire called the Barkley Adult ADHD Rating Scale, which measures inattention and hyperactivity. The second was the Quantitative behavior test. This computer-based assessment tracks head movement with an infrared camera while the user completes a repetitive visual task, yielding objective scores for activity, impulsivity, and inattention. After completing these initial tests, the participants began a standard daily regimen of long-acting methylphenidate. Following two months of this daily treatment, the researchers evaluated the participants again using the same questionnaire and computer test. They wanted to see if the behavioral changes recorded under the single initial dose aligned with the improvements observed at the two-month mark. [Add PsyPost to your preferred sources](https://www.google.com/preferences/source?q=psypost.org) The researchers first analyzed individual scores from the initial tests against the long-term results. Out of the original group, 45 participants completed the questionnaire during the single-dose phase. The data showed that early reductions in self-reported hyperactivity and impulsivity after a single dose corresponded with improvements in those exact areas two months later. However, changes in self-reported inattention under the single dose did not consistently predict long-term improvements in attention. This suggests that certain symptom domains are harder for patients to subjectively evaluate in the short term. Conversely, early changes in all three main scores on the computer-based behavior test mapped onto longer-term improvements. Participants who demonstrated better objective focus and less movement under the single dose were highly likely to show those same benefits after two months of daily medication. Next, the team used a machine learning technique known as Lasso regression to build predictive models. This specific algorithm automatically selects a limited set of the strongest predictive variables from a large pool of data, ignoring weaker factors to prevent the model from becoming overly complicated. This analysis combined the single-dose test results with baseline demographic and clinical traits. These baseline traits included age, intelligence quotient, years of education, and initial symptom severity. These combined models successfully predicted actual patient outcomes at the two-month follow-up for most of the tested measures. The models accurately forecast total symptom improvement, changes in hyperactivity, and shifts in computer-tracked impulsivity and inattention. The researchers noted that combining the single-dose response with baseline traits like intelligence and age improved the accuracy of the predictions beyond using the single-dose data alone. In a secondary analysis, the researchers looked closer at how different types of intelligence related to treatment outcomes. They found that visual-spatial problem-solving skills were linked to long-term improvements in attention. Meanwhile, verbal intelligence scores were associated with long-term reductions in hyperactivity and impulsivity. The study also found that older participants tended to see better improvements in impulsivity, while younger participants saw greater gains in attention. The study evaluated a specific demographic group, which shapes how the results apply to the broader population. The researchers exclusively enrolled adult men to limit potential biological variables in a related brain imaging analysis. As a result, the findings require validation in women and people of other genders to confirm whether the predictive tests work universally. Additionally, the participants had no co-occurring mental health or developmental conditions, such as autism. People with ADHD frequently have other diagnoses, which can alter how they respond to stimulant medications. Testing these predictive models in a more diverse patient group will help determine their utility in everyday medical settings. The models also need to be tested on entirely separate groups of patients. While the machine learning algorithms performed well on the data they were trained on, predictive tools often require external validation to ensure they do not just reflect the specific characteristics of the original sample. The single-dose prediction approach remains in the testing phase and is not yet a replacement for standard clinical monitoring. In the future, researchers hope to incorporate physiological measures, such as heart rate variability, to build even more accurate predictive profiles. Combining objective computer tests, questionnaires, and biological markers could eventually help doctors tailor prescriptions to individual patients from day one. The study, “[Clinical response to a single-dose methylphenidate challenge is indicative of treatment response at two months in adults with ADHD](https://doi.org/10.1038/s41398-025-03557-3),” was authored by Valeria Parlatini, Joaquim Radua, Hannah Thomas, Miguel Garcia-Argibay, Alessio Bellato, Samuele Cortese, and Declan Murphy.