Method and apparatus for estimating a cognitive decision made in response to a known stimulus from the corresponding single-event evoked cerebral potential

A method for estimating a cognitive decision made by a human in response to a known external stimulus, from the corresponding single-event, evoked cerebral potential, comprising the steps of:

• a) mapping the elements of a set of parameters uniquely one-to-one onto the elements of a set of internal cerebral potential sources for all possible decisions which can be made in response to the known external stimulus, where the value of the parameter mapping to the internal source corresponding to the true decision is unity and the values of the remaining parameters are zero;
• b) mapping said set of parameters uniquely one-to-one onto a set of attenuators of which the sum of the outputs from said attenuators is the deterministic input to an autoregressive filter used to model the single-event, evoked cerebral potential as a cerebral potential process, and the inputs to said attenuators are the internal cerebral potential sources, where the values of the said parameters are given by the attenuations for the corresponding sources;
• c) estimating the values of said parameters by an iterative process performed in turn for each of said possible internal sources as an initial estimate to the true source, by computing in sequence first the corresponding attenuation along with the coefficients of the autoregressive process from the recorded evoked cerebral potential and the waveform of the assumed source, and then computing an update to the source-waveform from the corresponding attenuation, with the product of said iterative process being a recursive sequence of estimations of the parameters for the internal cerebral sources;
• d) computing a reliability of the estimation from the human cognitive state by:
• (1) computing the power spectrum of the electroencephalogram from the autoregressive coefficients,
• (2) classifying the cognitive state of the human operator from the power spectrum components with embedded rules for state membership, and
• (3) computing the reliability of the estimated decision from the classified cognitive state with embedded fuzzy logic rules based on expert knowledge of the cognitive processes as a function of the state;
• and
• e) applying said sequence of estimations of the parameters and the estimation reliability as inputs to an artificial neural network used as a classifier, the outputs of which are a measure of the probabilities of the occurrences of all possible decisions which can be made in response to the known external stimulus, and selecting the strongest output of the classifier as the best estimate of the decision made.