PubMed Health⌕ Search

Biomedical subjects

R R Douglas

Publications and source records attributed to R R Douglas.

7 recordsLinked to original sources
PubMed1984

Comparative factor analysis models for an empirical study of EEG data, II: A data-guided resolution of the rotation indeterminacy.

In this paper (the second in a series), we consider a (generic) pair of datasets, which have been analyzed by the techniques of the previous paper. Thus, their "stable subspaces" have been established by comparative factor analysis. The pair of datasets must satisfy two confirmable conditions. The first is the "Inclusion Condition," which requires that the stable subspace of one of the datasets is nearly identical to a subspace of the other dataset's stable subspace. On the basis of that, we have assumed the pair to have similar generating signals, with stochastically independent generators. The second verifiable condition is that the (presumed same) generating signals have distinct ratios of variances for the two datasets. Under these conditions a small elaboration of some elementary linear algebra reduces the rotation problem to several eigenvalue-eigenvector problems. Finally, we emphasize that an analysis of each dataset by the method of Douglas and Rogers (1983) is an essential prerequisite for the useful application of the techniques in this paper. Nonempirical methods of estimating the number of factors simply will not suffice, as confirmed by simulations reported in the previous paper.

Animals↗
PubMed1984

Comparative factor analysis models for an empirical study of EEG data, III: Resolution of stationary generating signals with confirmation by variable deletion.

This paper (the third in a series), reports the results obtained from the application of two distinct methods for the estimation of the four generating signals of a specific, large EEG dataset, in the context of the factor analysis model. These four signals were "detected" (though not resolved) in the work reported in the first paper in this series, where a "stability computation" determined the four dimensional subspace spanned by the four desired signals. The second paper discussed the "signal detection" method of resolving generating signals, and this paper presents another method of resolution, called "variable deletion." The estimates of the four signals produced by these two techniques are in agreement with each other. Indirect evidence is presented that the EEG activity represented in terms of these four "signals" is generated by four stochastically independent random variables.

Cerebral Cortex↗
PubMed1983

Comparative factor analysis models for an empirical study of EEG data.

This paper (the first in a series) applies new, empirical factor analysis methods to the problem of "banding" EEG power spectra. A measure is introduced for the comparison of factor analysis results (factor loading matrices). The measure, Ambient Matrix Coherence (AC) is "geometrically unbiased", and invariant of so-called "oblique rotations." AC is used in a "stability computation" to find the dimension of the stable factor analysis solution common to several subsets of a given dataset. If the factor analysis model is appropriate, then the correct number of factors is empirically determined in this way. Stability computations were first performed on various simulated datasets to establish the robustness and efficacy of this method (for various noise levels). These techniques were then applied to EEG power spectra datasets for each of 8 leads. Comparison of these results indicated 3 stable factors in common to all 8 leads, an additional less stable factor in common to 5 leads, and weak stability for the six-dimensional solution for one lead.

Adult↗