EEG power spectra in normal and dyslexic children. I. Reliability during passive conditions.

Abstract Three minutes of passive eyes-closed (EC) and eyes-open (EO) EEG were recorded before and after 4–5 h of behavioral tasks in 10–12-year-old boys of normal intelligence and neurological status. Half were severely reading disabled; half were reading normally. Bilateral central, parietal, and mid-temporal EEG referenced both to vertex and to linked ears was recorded. Averaged FFT power spectra of artifact-free 1 sec epochs were computed within each recording condition. Test-retest reliabilities were computed for each band and for the entire spectrum using intra-class correlations (ICC). Reliabilities were assessed separately for each group, lead and reference condition, and for absolute power and relative power. The results reveal excellent reliability in the normal reading group. Reliabilities are higher with the Cz reference than the A1A2 reference, primarily in the delta and beta2 bands. Temporal recordings have lower reliabilities than central or parietal leads. These effects summate to yield poor reliabilities for the delta and beta2 bands for temporal-A1A2 recordings. Reliabilities for the dyslexic group are lower than control group values, yet are still acceptably high. Beta2 ICCs were markedly reduced in the dyslexics, possibly reflective of increased EMG artifact. Our finding that absolute power is a reliable as relative power is at odds with the report of John et al. (1980) that absolute power was not reliable enough to be useful in automated EEG assessment procedures. Use of absolute power is warranted whenever possible, since the interpretation of findings based solely upon relative power can be ambiguous. Our results indicate that under properly controlled conditions, excellent reliability of both absolute and relative power even for the passive EC and EO conditions can be obtained. These findings support the utility of EEG power spectra as a reliable index of brain function for studies of normal and learning disabled children.

[1]  M. Matousek,et al.  Serial quantitative electroencephalography. , 1979, Electroencephalography and clinical neurophysiology.

[2]  H. Stassen,et al.  Computerized recognition of persons by EEG spectral patterns , 1980 .

[3]  D. Galin,et al.  Dyslexic children have normal vestibular responses to rotation. , 1983, Archives of neurology.

[4]  L. Lindholm,et al.  Alpha amplitude dependence on skull thickness as measured by ultrasound technique. , 1970, Electroencephalography and clinical neurophysiology.

[5]  H. Ahrens,et al.  Multivariate variance‐covariance components (MVCC) and generalized intraclass correlation coefficient (GICC) , 1976 .

[6]  G. T. Duncan,et al.  A Monte-Carlo study of asymptotically robust tests for correlation coefficients , 1973 .

[7]  M. Corner,et al.  Individual differences in the human electroencephalogram during quiet wakefulness. , 1979, Electroencephalography and clinical neurophysiology.

[8]  B. J. Winer.  Statistical Principles in Experimental Design , 1992 .

[9]  M. Kennard,et al.  A longitudinal study of electroencephalographic frequency patterns in mental hospital patients and normal controls. , 1957, Electroencephalography and clinical neurophysiology.

[10]  Helmer R. Myklebust.  Progress in Learning Disabilities , 1982 .

[11]  H. Kraemer.  On estimation and hypothesis testing problems for correlation coefficients , 1975 .

[12]  E. John,et al.  Developmental equations for the electroencephalogram. , 1980, Science.

[13]  D. O. Walter,et al.  Temporal stability and individual differences in the human EEG: an analysis of variance of spectral values. , 1968, IEEE transactions on bio-medical engineering.