Version 1

Objectives: 1to evaluate the ability of the power spectrum of brain electrical activity to discriminate between age matched normal children and children diagnosed as learning disabled (LD) and, 2to correlate brain electrical activity with intelligence, school achievement and neuropsychological performance test scores. Methods: Electrical potentials (EEG) were recorded from 19 channels in a resting eyes closed state for 1 to 10 minutes in two groups of children (normal, N= 326 & LD, N = 66). The normal group of children were age matched to the LD group and randomly divided into two sub-groups: a test group (N = 163) and a replication group (N = 163). Electrophysiological spectral measurements were compared between the LD and test normal control groups by t-tests and then variables with P <.05 were factor analyzed. A total of 61 variables were selected and entered into a discriminant analysis based on being the two highest loading variables on each of five factors with eigenvalues > 1. Cross-validation of the discriminant function involved the independent classification of the second normal group of subjects as well as by correlations between behavioral variables such as the Wide range Achievement Test and WISC-R sub-tests and the EEG discriminant scores. Results: The test discriminant accuracy = 86.34% and the validation discriminant was 83.2% accurate. The order of effect size was EEG phase delays > EEG coherence > EEG amplitude asymmetries > absolute power > relative power. The multiple regression between electrophysiology and neuropsychological and school performance was typically in the range of r = 0.5 to r = 0.75 (P < .000001). Conclusions: EEG network measures such as coherence and phase delays were stronger discriminating variables than were power measures. The accuracy of the discriminant function was similar to that reported in other studies and supports the view that learning disabilities is a heterogeneous population. The importance of the discriminant analysis is in providing a deeper understanding of the underlying neurophysiological correlates of learning disabilities.