‘Serial magnetoencephalographic observations of a normally developing infant during sleep’
S. Bowyer,D. Burdette,J. Moran,N. Tepley
DOI: https://doi.org/10.1111/j.1469-8749.2000.tb00059.x
2000-02-01
Developmental Medicine & Child Neurology
Abstract:‘Serial magnetoencephalographic observations of a normally developing infant during sleep’ SIR–Magnetoencephalography (MEG) is the measurement of weak magnetic fields arising from neuronal activity in the brain. MEG has been used extensively in the evaluation of functional localization1, epileptiform activity2, and DC shifts, which arise in migraine, stroke, and other pathological conditions3–5. With the advent of large array neuromagnetometers, specifically those utilizing magnetometers that can detect relatively deeper sources, infants are easier to study. EEG has been used in the assessment of brain maturity in infants and young children. Although the waking background provides a useful marker of maturity (e.g. frequency of the posterior, dominant rhythm), waking EEG recordings in the infant are often limited by prominent muscle and movement artefacts. Sleep EEG recording provides a more consistent means of assessing infantile brain maturity6. As the brain matures, normal myelination produces an increase in faster frequencies and a decrease in the duration and amplitude of sleep spindles in EEG6,7. Weekly magnetoencephalographic measurements were performed on a normally developing infant during a longitudinal study of 47 weeks. This feasibility study was conducted to evaluate changes in brain magnetocerebral fields that accompany normal maturation in the infant. MEG recordings were made with a seven-channel superconducting neuromagnetometer (BTi model 607). MEG signals were recorded first over the left hemisphere and then over the right during each recording session. The central coil in the seven-channel array was located over the midtemporal region in the T3 position on the left (International 10–20 System of electrode placement) and the T4 position on the right. Six minutes of MEG data were collected over each hemisphere at a rate of 130 samples per channel per second, with low pass filter set at 50 Hz and high pass filter set at 0.5 Hz. Data were stored on optical disk for later analysis. A single power spectrum estimate for each 6 minutes of data was generated by averaging fast fourier transform (FFT) power spectra, which utilized an 8-second Hanning window and overlapped by 75%. To avoid amplitude changes from variations of the sensor-to-head distance, these spectral amplitudes were converted into a percentage of the total spectral power, a measure relatively independent of distance. As sleep spindles appear as early as 6 to 8 weeks postterm, we analyzed the spectral power of sleep spindles in this normally developing infant. Frontally dominant 10to 12-Hz spindles7 and central maximum 14-Hz spindles were analyzed separately. Spectral power within delta, theta, alpha, and beta bands was analyzed. Figure 1a shows the significant decrease in the theta power band observed over the 38 weeks, and Figure 1b shows a similar increase in the alpha power band. There was no significant change in the beta power, and the delta power bands were obscured by vibration artifact from the probe and gantry in the 2to 3Hz range. Sleep spindle bursts for week 19 lasted 1.6 seconds,