EEG Biomarkers for ADHD Stimulant Treatment

Pediatric attention deficit hyperactivity disorder (ADHD) affects up to 10% of children in the U.S. and more than 90% are prescribed stimulant medications according to clinical guidelines. The standard of care for pharmacological treatment of ADHD is a "trial-and-error" approach that requires frequent titration, side effects management, and communication among providers, parents, and school personnel over weeks, months, and years. This level of effort often exceeds the capacity of primary care providers, who manage treatment for the majority of affected children. In the first year following stimulant prescription, recommended routine follow-up occurs in <50% of cases treated in primary care. Consequently, suboptimal dosing and early discontinuation of stimulants are exceedingly common, with >75% pediatric ADHD patients showing suboptimal medication adherence in the first year.

The current proposal will utilize electroencephalography (EEG), a non-invasive, affordable brain measurement tool, to characterize predictive biomarkers for response to two commonly prescribed stimulant medication classes: methylphenidate (MPH) and amphetamines (AMP). Predictive biomarkers can be used to guide clinical decision making and treatment selection for pediatric ADHD, ultimately shortening the time from diagnosis to treatment optimization; reducing burden on clinicians, families, and teachers; and leading to better lifetime outcomes for children with ADHD. Thus far, behavioral- and genomic-based treatment biomarkers have not been successful for ADHD, and there is increasing demand by patients and providers to identify a solution.

Building on our pilot data, this study will constitute a prospective, single-blind, double-baseline randomized MPH-AMP crossover trial with 220 treatment-naïve, 7-11-year-old children with ADHD. EEG will be collected prior to each medication arm, and treatment response will be determined using 3-week rapid titration protocols, consistent with our pilot study and other published trials. Trained clinicians will characterize each participant as a "responder" or "non-responder" for each medication.

EEG features with small to medium effect sizes that are predictive of positive response to MPH and AMP will be identified. Preliminary data suggest MPH and AMP will have distinct EEG predictive biomarkers. To further examine the translational potential of these results, test-retest reliability and classification agreement of biomarkers over 24 hours and 4 weeks will be estimated. Results of this study have potential to accelerate pediatric mental health treatment development and precision medicine care for ADHD.

• methylphenidate HCl (Drug): 3-week titration of liquid Quillivant XR
• Mixed amphetamine salts extended release (Drug): 3-week titration of liquid Dyanavel XR

Primary outcomes

• Medication A Clinical Global Impressions - Improvement Scale (CGI-I) [Week 3 Virtual Visit (Day 22)]: Clinician rating of participant ADHD symptom improvement on medication A. The Clinical Global Impressions - Improvement Scale (CGI-I) is rated by a trained study clinician on a scale of 1-7, with scores of 1 ("very much improved") and 2 ("much improved") corresponding to clinically significant symptom improvement. Scores of 3-7 will indicate no improvement or worsening of symptoms.
• Medication B Clinical Global Impressions - Improvement Scale (CGI-I) [End of Study Virtual Visit (Day 50)]: Clinician rating of participant ADHD symptom improvement on medication B. The Clinical Global Impressions - Improvement Scale (CGI-I) is rated by a trained study clinician on a scale of 1-7, with scores of 1 ("very much improved") and 2 ("much improved") corresponding to clinically significant symptom improvement. Scores of 3-7 will indicate no improvement or worsening of symptoms.
• Frontal Theta Beta Ratio [Eligibility Visit (Day 0), Baseline Visit (Day 1), Crossover Visit (Day 29)]: Theta-Beta Ratio (TBR): TBR will be calculated from periodic resting EEG. Periodic power isolates the oscillatory component of the EEG by subtracting the aperiodic power spectra from the total power. Periodic power is extracted with the Fitting Oscillations and One-Over-f (FOOOF) MATLAB toolbox which parametrizes resting state EEG data into aperiodic and periodic components. Resting spectral power values will be averaged over theta (4-6 Hz) and beta (13-30 Hz) frequencies across midline electrode clusters (frontal: Fz, F3, F4; central: Cz, C3, C4; parietal: Pz, P3, P4; occipital: Oz, O1, O2). Our primary analyses will focus on frontal TBR during the Lights-Off Resting condition.
• Aperiodic Dynamics [Eligibility Visit (Day 0), Baseline Visit (Day 1), Crossover Visit (Day 29)]: Aperiodic Dynamics: The aperiodic exponent for lights-off and lights-on resting conditions will be averaged across midline electrode clusters FOOOF MATLAB toolbox. Aperiodic dynamic values will be computed for each participant as the difference between lights-on and lights-off exponent, divided by the lights-off resting exponent, as in prior publications. Thus, higher scores will indicate greater exponents in the lights-on experiment. Primary analyses will use the central electrode cluster.
• P100 Visual Evoked Potential Amplitude [Eligibility Visit (Day 0), Baseline Visit (Day 1), Crossover Visit (Day 29)]: VEP P100 Amplitudes. The P100 VEP component will be derived from a passive pattern-reversal visual evoked potential (VEP) task that includes 200 500ms trials of a black and white checkerboard reversal over 3 minutes. P100 amplitudes will be averaged over an occipital electrode cluster (Oz, O1, O2) from approximately 75-150ms and extracted for each trial. Primary analyses will focus on mean VEP amplitude over trials.
• Vanderbilt Rating Scales [Baseline, Medication A Week 1, Medication A Week 2, Medication A Week 3, Crossover Week 4, Medication B Week 5, Medication B Week 6, End of Study Week 7]: The Vanderbilt Rating Scales include 18 items corresonding to DSM-5 ADHD symptoms. Parents and teachers will each complete the scales using an electronic form. Items are rated on a scale of 0-3, with 0 = symptom never or rarely occurs, and 3 = symptom often or almost always occurs. Total symptom severity is calculated as the sum of all 18 items; inattention symptom severity is calculated as the sum of items 1-9; hyperactivity/impulsivity symptom severity is calculated as the sum of items 10-18.

Secondary outcomes

• P100 Visual Evoked Potential Habituation [Eligibility Visit (Day 0), Baseline Visit (Day 1), Crossover Visit (Day 29)]: VEP P100 Amplitudes. The P100 VEP component will be derived from a passive pattern-reversal visual evoked potential (VEP) task that includes 200 500ms trials of a black and white checkerboard reversal over 3 minutes. P100 amplitudes will be averaged over an occipital electrode cluster (Oz, O1, O2) from approximately 75-150ms and extracted for each trial. Secondary analyses will focus on P100 amplitude changes over trial, i.e. habituation coefficients.
• Side Effects Rating Scale, Parent- and Teacher-Report [Baseline, Medication A Week 1, Medication A Week 2, Medication A Week 3, Crossover Week 4, Medication B Week 5, Medication B Week 6, End of Study Week 7]: Adverse effects associated with stimulant treatment will be recorded by parents and teachers on a weekly basis on the Vanderbilt Follow-Up rating scales, which include an adapted version of the Pittsburgh Side Effects Rating Scale. 12 side effects symptoms are rated on a scale of 0 = not currently a problem to 3 = severe problem. Side effects will be used for clinical decision making. Secondary exploratory analysis will examine associations between EEG features and total severity of side effects by summing ratings from the 12 items.
• EEG Coherence [Eligibility Visit (Day 0), Baseline Visit (Day 1), Crossover Visit (Day 29)]: EEG coherence will be derived from resting state EEG as well as ERP tasks. Both global (i.e., long-range) and local (i.e., within region of interest) coherence will be examined in exploratory analyses.
• N2 Event Related Potential Amplitude [Eligibility Visit (Day 0), Baseline Visit (Day 1), Crossover Visit (Day 29)]: The HAPPE+ER pipeline will be used to extract N2 amplitudes during the Go-NoGo task, by condition and electrode. Mean amplitudes will be calculated for components across regions of interest, following visual inspection of temporal and topographic plots.
• P300 Event Related Potential Amplitude [Eligibility Visit (Day 0), Baseline Visit (Day 1), Crossover Visit (Day 29)]: The HAPPE+ER pipeline will be used to extract P300 amplitudes during the Go-NoGo task, by condition and electrode. Mean amplitudes will be calculated for components across regions of interest, following visual inspection of temporal and topographic plots.

Sex: All, ages: 7 Years to 11 Years

Inclusion Criteria:

1. Ages 7:0 - 10:11 (years:months)
2. Has a diagnosis of ADHD or being evaluated for ADHD
3. Stimulant naïve or previously trialed stimulant medications for < 6 months without achieving symptom remission, per caregiver report and/or medical chart review (if available)
4. CGI-Severity rating of 4 "Moderately ill" through 6 "Severely ill."
5. Willing and able to comply with study procedures

Exclusion Criteria:

1. Use of stimulants or other psychotropic medications within 7 days before Eligibility Visit
2. History of severe side effects to stimulants (suicidality, complete loss of appetite, cardiopulmonary complications) per caregiver report or medical chart review, determined by the study MD
3. Intellectual disability or IQ < 75 per medical chart review or performance on standardized cognitive testing during the Eligibility Visit
4. Diagnosis of Autism spectrum disorder (ASD) per medical chart review or caregiver report
5. Fetal alcohol exposure per medical chart review or caregiver report
6. Current suicidal ideation per caregiver or child report on CSSRS
7. Non-febrile seizures per caregiver report or medical chart review
8. Cardiopulmonary conditions, pregnancy or other medical conditions that contraindicate psychostimulant use per determination by the study MD