Análisis avanzado de señales de potenciales evocados multifocales aplicados al diagnóstico de neuropatías ópticas

Zusammenfassung

Objective: The objective of this thesis is to propose new signal processing methods applied to multifocal visual evoked potentials (mfVEP) technique. These methods will improve the objective evaluation of the visual pathway in multiple sclerosis risk patients and enhanced the applicability of this technique for the clinical practice. Patients: Four cohorts groups have been used: 22 controls subject, 15 radiologically isolated syndrome, 28 clinical isolated syndrome and 28 definitive multiple sclerosis patients. Eyes from these patients have been divided in affected/no affected by optic neuritis. Methods: Three studies have been carried out in this thesis. First, amplitudes from mfVEP recordings have been analyzed according to concentric rings. SNR of each ring and significance difference between subjects were computed. Eyes were classified as normal/abnormal mfVEP amplitude according a threshold criterion, and these eyes were correlated with the disability level of each patient. In the second study a new method to compute latency progression was developed based on the comparison between signal from the same patients recorded at different times. In the third study the Empirical Mode Decomposition was applied to the analysis of mfVEP amplitudes and to the interocular latencies computing. Results: First study found that mfVEP amplitude significatively decrease when the risk of suffering EM is increased, especially in 3 and 5 rings. These rings also present lower errors when the eyes are classified according the amplitude and also present highest correlation with disability scale. The latency computation method presented in the second study achieved values close to the ideal values with less variability and highest values of not analyzable sectors than the standard method. In the third study, differences in amplitude between the patient’s groups are magnified when EMD is applied to the register (even more in the ring 5). Less variability has been achieved in interocular latencies computations when EMD is applied to the recordings. Conclusions. This thesis presents new signal processing methods applied to mfVEP. These methods could improve the detection of patients in MS-risk and quantify the progression of the disease.