Independent components in stimulus-related BOLD signals and estimation of the underlying neural responses.

We measured blood oxygen level dependent (BOLD) responses to the onset of dynamic noise stimulation in defined regions of the primary retinotopic projection (V1) in visual cortex. The response waveforms showed a remarkable diversity across stimulus types, violating the basic assumption of a unitary general linear model of a uniform BOLD response function convolved with each stimulus sequence. We used independent component analysis (ICA) to analyze the component mechanisms contributing to these responses. The underlying neural responses for the components were estimated by nonlinear optimization through the Friston-Buxton hemodynamic model of the BOLD response. Our analysis suggests that one of the identified components reflected a sustained neural response to the stimulus and that another reflected an extremely slow neural response. A third component exhibited nonlinear change-specific transient responses. The first two components showed stable spatial structure in the V1 region of interest with respect to the eccentricity of the noise stimulus.

Electrophysiological measures of acute cerebral ischaemia.

A method of EEG analysis is described which provides new insights into EEG pathology in cerebral ischaemia. The method is based on a variant of detrended fluctuation analysis (DFA), which reduces short (10 s) segments of spontaneous EEG time series to two dimensionless scaling exponents. The spatial variability of each exponent is expressed in terms of its statistical moments across EEG channels. Linear discriminant analysis combines the moments into concise indices, which distinguish normal and stroke groups remarkably well. On average over the scalp, stroke patients have larger fluctuations on the longest time scales. This is consistent with the notion of EEG slowing, but extends that notion to a wider range of time scales. The higher moments show that stroke patients have markedly reduced variability over the scalp. This contradicts the notion of a purely focal EEG scalp topography and argues instead for a highly distributed effect. In these indices, subacute patients appear further from normal than acute patients.

Weighted regularization in electrical impedance tomography with applications to acute cerebral stroke.

We apply electrical impedance tomography to detect and localize brain impedance changes associated with stroke. Forward solutions are computed using the finite-element method in two dimensions. We assume that baseline conductivity values are known for the major head tissues, and focus on changes in the brain compartment only. We use singular-value decomposition (SVD) to show that different impedance measurement patterns, which are theoretically equivalent by the reciprocity theorem, have different sensitivities to the brain compartment in the presence of measurement noise. The inverse problem is solved in part by standard means, using iterated SVD, and regularizing by truncation. To improve regularization we introduce a weighting scheme which normalizes the sensitivity matrix for voxels at different depths. This increases the number of linearly independent components which contribute to the solution, and forces the different measurement patterns to have similar sensitivity. When applied to stroke, this weighted regularization improves image quality overall.

Scalp electrode impedance, infection risk, and EEG data quality.

OBJECTIVES: Breaking the skin when applying scalp electroencephalographic (EEG) electrodes creates the risk of infection from blood-born pathogens such as HIV, Hepatitis-C, and Creutzfeldt-Jacob Disease. Modern engineering principles suggest that excellent EEG signals can be collected with high scalp impedance ( approximately 40 kOmega) without scalp abrasion. The present study was designed to evaluate the effect of electrode-scalp impedance on EEG data quality. METHODS: The first section of the paper reviews electrophysiological recording with modern high input-impedance differential amplifiers and subject isolation, and explains how scalp-electrode impedance influences EEG signal amplitude and power line noise. The second section of the paper presents an experimental study of EEG data quality as a function of scalp-electrode impedance for the standard frequency bands in EEG and event-related potential (ERP) recordings and for 60 Hz noise. RESULTS: There was no significant amplitude change in any EEG frequency bands as scalp-electrode impedance increased from less than 10 kOmega (abraded skin) to 40 kOmega (intact skin). 60 Hz was nearly independent of impedance mismatch, suggesting that capacitively coupled noise appearing differentially across mismatched electrode impedances did not contribute substantially to the observed 60 Hz noise levels. CONCLUSIONS: With modern high input-impedance amplifiers and accurate digital filters for power line noise, high-quality EEG can be recorded without skin abrasion.

Regional head tissue conductivity estimation for improved EEG analysis.

We develop a method for estimating regional head tissue conductivities in vivo, by injecting small (1-10 microA) electric currents into the scalp, and measuring the potentials at the remaining electrodes of a dense-array electroencephalography net. We first derive analytic expressions for the potentials generated by scalp current injection in a four-sphere model of the human head. We then use a multistart downhill simplex algorithm to find regional tissue conductivities which minimize the error between measured and computed scalp potentials. Two error functions are studied, with similar results. The results show that, despite the low skull conductivity and expected shunting by the scalp, all four regional conductivities can be determined to within a few percent error. The method is robust to the noise levels expected in practice. To obtain accurate results the cerebrospinal fluid must be included in the forward solution, but may be treated as a known parameter in the inverse solution.

Computational rules for chemotaxis in the nematode C. elegans.

We derive a linear neural network model of the chemotaxis control circuit in the nematode Caenorhabditis elegans and demonstrate that this model is capable of producing nematodelike chemotaxis. By expanding the analytic solution for the network output in time-derivatives of the network input, we extract simple computational rules that reveal how the model network controls chemotaxis. Based on these rules we find that optimized linear networks typically control chemotaxis by computing the first time-derivative of the chemical concentration and modulating the body turning rate in response to this derivative. We argue that this is consistent with behavioral studies and a plausible mechanism for at least one component of chemotaxis in real nematodes.