EEG spectral parameters were computed in a group of children with different degrees of difficulty in learning to read and write. For statistical analyses, Z-transformed values according to normative age-regression equations were used to control the age effects. Canonical Correlation Analysis between absolute power in different bands and the categories of the educational evaluation (good, regular, poor and very poor) showed that more delta was probably related to a poor evaluation and more alpha in occipital areas to a good one. MONOVA also showed highly significant differences in the absolute power in many leads between children with different evaluations. As children with a poor evaluation very frequently had antecedents of risk factors related to brain damage and were from a low socioeconomic status, and both factors have been shown to affect absolute power, it may be that the differences observed were due to these causes. However, relative power correlated more with the learning problems. Children with minor difficulties, with no antecedents and with good socioeconomic status had more theta in almost all leads than children with a good evaluation and with the same characteristics. Children with a poor, or very poor, evaluation had more delta in left frontal and temporal areas (F3, F7 and T3) which may reflect underlying cerebral dysfunction of these regions directly involved in reading and writing processes.

No abstract

It is shown that hidden Markov models (HMMs) are a powerful tool in the analysis of multielectrode data. This is demonstrated for a 30-electrode measurement of neuronal spike activity in the monkey's visual cortex during the application of different visual stimuli. HMMs with optimized parameters code the information contained in the spatiotemporal discharge patterns as a probabilistic function of a Markov process and thus provide abstract dynamical models of the pattern-generating process. We compare HMMs obtained from vector-quantized data with models in which parametrized output processes such as multivariate Poisson or binomial distributions are assumed. In the latter cases the visual stimuli are recognized at rates of more than 90% from the neuronal spike patterns. An analysis of the models obtained reveals important aspects of the coding of information in the brain. For example, we identify relevant time scales and characterize the degree and nature of the spatiotemporal variations on these scales.

No abstract

No abstract