Neuraxis
ADHD Decision Support
EEG-based attention and hyperactivity biomarker analysis for ADHD decision support. The most advanced module in the SynapCore platform, with preliminary research results in a paediatric cohort. Adult results are under revalidation.
What Neuraxis Does
Neuraxis analyses EEG recordings of children aged 7 to 12 for patterns associated with attention-deficit/hyperactivity disorder (ADHD). It runs the recording through SynapCore's pipeline, extracts spectral, coherence and complexity features, and produces a calibrated probability that the recording shows ADHD-consistent patterns, together with the features that drove it.
The output is a structured decision-support report designed for review by a qualified clinician. Neuraxis does not diagnose. It adds EEG-based information to the clinical assessment.
Spectral Band Features
Of all the features Neuraxis uses, these three frequency bands are the ones most studied in ADHD research.
Theta (4-8 Hz)
Higher theta power, especially over frontal regions, is one of the most often reported EEG findings in ADHD research. Neuraxis measures absolute and relative theta power at every electrode.
Alpha (8-13 Hz)
Alpha activity reflects cortical arousal and attentional state. Neuraxis measures alpha power and the alpha/theta ratio at every electrode.
Beta (13-30 Hz)
Beta activity is linked to active cognitive processing and motor inhibition. Neuraxis measures beta power and the theta/beta ratio (TBR), which has been widely studied as a possible ADHD marker.
LightGBM Classifier
Neuraxis uses a LightGBM gradient-boosted decision tree model, chosen because it works well on tabular EEG features with small cohorts and its predictions can be explained.
LOSO Cross-Validation
All metrics come from leave-one-subject-out cross-validation. Each subject is held out completely during training, so the model is always tested on a person it has not seen.
Calibrated Probabilities
The raw LightGBM score passes through a sigmoid calibrator fitted on out-of-fold predictions, so the reported probability is closer to an observed rate on the training cohort. It has not been recalibrated on any clinical cohort.
Feature Importance
Each prediction is explained with SHAP values computed on the trained LightGBM model, so a clinician can see which EEG features pushed the probability up or down for that recording.
Performance Metrics
All results come from LOSO cross-validation on public research datasets. They are preliminary research metrics, not clinical performance claims.
Neuraxis-Child v1.0
Prototype · Research Use OnlyAUC
0.650
Balanced Accuracy
0.595
Validation
LOSO CV
Dataset
IEEE ADHD
Population
Paediatric
Age Range
7-12 years
Neuraxis-Adult v1.0
Under revalidation · Research Use OnlyAUC
Not reported
Balanced Accuracy
Not reported
Validation
LOSO CV
Dataset
TDBRAIN
Population
Adult
Age Range
18-35 years
Internal LOSO results on the IEEE cohort. On an independent external cohort (HBN-EEG, N=321, ages 5 to 17) Neuraxis-Child scored AUC 0.519, which is chance level. All reported metrics are preliminary research results. Research Use Only (RUO).
Research Datasets
IEEE ADHD Dataset
Publicly available paediatric EEG dataset with clinical ADHD diagnoses. Used as the primary training and validation set for Neuraxis-Child v1.0.
TDBRAIN
Lifespan EEG database (ages 5 to 89) from the Brainclinics Foundation, Nijmegen, The Netherlands (van Dijk et al., 2022). Our subset of 72 subjects covers the adult ADHD age range and was recorded with different equipment from the IEEE data. Used in Neuraxis-Adult validation work; the adult module is under revalidation.
Research Use Only
Neuraxis is in research and development. It is not approved, cleared or certified for clinical diagnostic use by any regulatory authority. All results are preliminary research metrics obtained on public datasets. Neuraxis does not diagnose ADHD. It is designed to support qualified clinicians.