Élaboration d'un outil logiciel d'analyse des données de spectroscopie proche infrarouge pour l'étude du vieillissement cognitif

The community of neuroscientists in cognition and aging has always enthusiastically welcome new cerebral imaging modalities. Previously, the brain had been studied only through the performance of subjects while carrying out cognitive tasks. Since then, imaging techniques have allowed to literally see the cortical activity. However, the deep rooted dependence inbetween nervous and circulatory systems makes the interpretation of imaging data difficult. The capability of continuous wave near infrared spectroscopy (continuous wave NIRS) to follow both oxyhemoglobin (HbO) and deoxyhemoglobin (HbR) concentrations has offered an essential investigation mean into the study of the brain neurophysiology.

In the studies of cognitive aging we aim to compare cerebral activities amongst population of different ages. Concerning optical imaging, continuous wave NIRS signal is affected by variations in anatomy as well as optical properties of the tissues. Thus any changes related to aging have an impact on our observations through the continuous wave NIRS. One of the important factors in models describing aging is the differential path length factor (DPF). The latter corrects the bias introduced to the observations by the combination of anatomical and physiological variations with age. Whereas studies over senior population have had a growing interest, this factor has not been evaluated for individuals aged over 51.

It is from the necessity of such a correction that emerged the essential objective of this project. Its ambitious aim is to provide one of the first toolboxes with the promise of giving more clarification to continuous wave NIRS signal by normalizing the detected activities in the cortex. An analysis pipeline of multimodal acquired data from magnetic resonance imaging (MRI), both anatomical and functional, time resolved spectroscopy (TRS) and continuous wave NIRS has been designed and integrated into the statistical analysis software Statistical Parametric Mapping (SPM) developed by the Wellcome Trust Centre for Neuroimaging (UCL). Its application in cerebral imaging permits the processing of large amounts of data quite easily and gives statistical map of cortical activations. Thanks to the new horizons that the integrated NIRS toolbox in SPM offers, a promising finding was achieved concerning the DPF; for the first time, it was evaluated up to the age of 72.

The results of this study show that at the fontal cortex, simulations over anatomical MR images applying TRS measures don’t prove any differentiation of the DPF with age. However, our results are underestimated systematically with respect to DPF values published in 1996 by Duncan et al. and vastly quoted since then. Our conclusion would be that the absorption and diffusion properties (μ_a and μ _s respectively) don’t correspond to the intrinsic characteristics of the brain tissues. In other words, the DPF would depend on μ _a and μ_s more than on the anatomy of the media where the light travels. Thereby, the individualization of μ _a and μ_s properties would be definitely a decisive step towards reliable reconstructions of cerebral activity maps.