Biomedical subjects

Andrew C James

Publications and source records attributed to Andrew C James.

4 recordsLinked to original sources

PubMed2005

Sparse multifocal stimuli for the detection of multiple sclerosis.

We compared the diagnostic capabilities of contrast reversal and sparse pattern pulse stimulation for dichoptic multifocal visual evoked potentials (mfVEPs) measured in normal subjects and multiple sclerosis (MS) patients. Multifocal responses were obtained from 27 normal subjects and 50 relapsing-remitting patients, 26 of whom had experienced optic neuritis (ON+). The patient groups were matched for length of disease and number of clinical attacks. Compared with the responses of normal subjects those of MS patients had significantly smaller response amplitudes, lower signal-to-noise ratios, more complex response waveforms, and longer response delays. The effects were larger for sparser stimuli. Sensitivities and specificities for the different stimulus types were estimated from receiver operator characteristic (ROC) plots. Bootstrap estimates of the accuracies of the ROCs for the most promising measure, the template delays, indicated the sparsest stimulus would deliver 92% sensitivity at a false-positive rate of 0%. In contrast, at 92% sensitivity the conventional mfVEP stimulus misdiagnosed more than 20% of the normal population. The results were similar for patients with no history of ON (ON-). In performing well in patients with no history of ON, the sparse mfVEPs seem to measure progressive damage associated with axon and gray matter losses rather than damage associated with a history of serious inflammation.

Adult↗

PubMed2004

Sensitivity and specificity of five abundance estimators for high-density oligonucleotide microarrays.

MOTIVATION: A number of algorithms have been proposed for the processing of feature-level data from high-density oligonucleotide microarrays to give estimates of transcript abundance. Performance in the common task of detecting differential expression between samples can be quantified by the statistical concepts of sensitivity and specificity, and represented by the use of receiver operating characteristic curves. These have been previously presented for small numbers of genes known to be differentially present in spiked-in samples. We present here a study of performance over a large number (thousands) of transcripts for which there is strong evidence of differential expression, with corresponding false positive rates controlled by comparisons between replicates. RESULTS: The straight-line regression analysis of a mixture series with replicates by five estimation algorithms produces a consensus set of 4462 transcripts with differential expression of agreed direction and high significance (p < 0.01) according to all algorithms. The more difficult task of two-sample tests between adjacent mixture levels produces performance curves of fraction true positive detected against significance level. Performance varies significantly between algorithms: at the p < 0.01 level, the detection rate varies between 41 and 66%. A control using comparisons between replicates at the same levels indicates that the tests produce empirical false positive rates closely matching the nominal p-values.

Algorithms↗

PubMedDate not supplied

Effect of temporal sparseness and dichoptic presentation on multifocal visual evoked potentials.

Multifocal VEP (mfVEP) responses were obtained from 13 normal human subjects for nine test conditions, covering three viewing conditions (dichoptic and left and right monocular), and three different temporal stimulation forms (rapid contrast reversal, rapid pattern pulse presentation, and slow pattern pulse presentation). The rapid contrast reversal stimulus had pseudorandomized reversals of checkerboards in each visual field region at a mean rate of 25 reversals/s, similar to most mfVEP studies to date. The rapid pattern pulse presentation had pseudorandomized presentations of a checkerboard for one frame, interspersed with uniform grey frames, with a mean rate of 25 presentations/s per region per eye. The slow pattern pulse stimulus had six presentations/s per region per eye. Recording time was 5.3 min/condition. For dichoptic presentation slow pattern pulse responses were 4.6 times larger in amplitude than the contrast reversal responses. Binocular suppression was greatest for the contrast reversal stimulus. Consideration of the signal-to-noise ratios indicated that to achieve a given level of reliability, slow pattern pulse stimuli would require half the recording time of contrast reversal stimuli for monocular viewing, and 0.4 times the recording time for dichoptically presented stimuli. About half the responses to the slow pattern pulse stimuli had peak value exceeding five times their estimated standard error. Responses were about 20% smaller in the upper visual field locations. Space-time decomposition showed that responses to slow pattern pulse were more consistent across visual field locations. We conclude that the pattern pulse stimuli, which we term temporally sparse, maintain the visual system in a high contrast gain state. This more than compensates for the smaller number of presentations in the run, and provides signal-to-noise advantages that may be valuable in clinical application.

Adult↗

PubMedDate not supplied

Hierarchical decomposition of dichoptic multifocal visual evoked potentials.

Visual evoked responses to dichoptically presented multifocal stimuli were recorded for 92 eyes. Two stimulus variants were explored: temporally sparse and rapidly contrast reversing. We used hierarchical decomposition (HD) to represent the multifocal responses in terms of a small number of potentially unique component waveforms that are interrelated in a multivariate linear autoregressive (MLAR) relationship. The HD method exploits temporal correlations over a range of delays in the responses to estimate parallel, feedforward and feedback relationships between the HD components. Three HD components having temporal interrelationships constrained (at P < 0.05) to a moving approximately 20 ms window could describe the multifocal responses well (median r2-values up to 90%). HD components were similar for both stimulus types and the component waveforms were temporally correlated, especially the first and third components. The data set was large enough to estimate separate HD components for each multifocal stimulus region. The component waveforms differed somewhat by region but the MLAR relationships were similar. At short delays parallel processing dominated. At longer delays the proportion of response drives that were attributed to feedback and feedforward relationships grew. Overall HD analysis seems to provide an informed summary of multifocal responses and insights into their sources.

Adult↗