| |
 |
UNIVERSITY OF BUCHAREST
FACULTY OF PHYSICS Guest
2024-11-22 1:36 |
 |
|
|
|
Conference: Bucharest University Faculty of Physics 2007 Meeting
Section: Electricity and Biophysics
Title:
Exploring Human Mind
Authors:
Radu Mutihac
Affiliation:
University of Bucharest, Physics Department
E-mail
mutihac@astralnet.ro
Keywords:
Functional brain imaging, wavelets, confirmatory and exploratory analysis, statistical parametric maps.
Abstract:
From such basics as breathing to the complexity of emotions and personality, every aspect of human life is governed by the brain. The key to the brain’s power and versatility resides in two general principles of cerebral function: functional specialization of brain regions (different brain regions perform different tasks) and functional integration (cerebral functions are carried out by networks of distinct interacting regions). Due to their complementarity, it is often necessary to employ both in order to resolve all the functional components of a given cerebral process.
The goal in brain imaging is to maximally suppress noise along with preserving as much as possible of the image features. Analysis methods, which reveal statistical regularities in data associated with brain function, can be loosely dichotomized in: (i) hypothesis-driven (confirmatory) and (ii) data-driven (exploratory) analysis. Most of imaging neuroscience relies on confirmatory inferential analysis, which makes use of spatially extended processes like statistical parametric mapping. The expected changes are specified as regressors in a multiple linear regression framework and the estimated regression coefficients are tested against a null hypothesis. The voxel-wise test statistics form summary images (maps) that are representations of the spatial distribution of functional activity induced by the experimental task. In contrast, exploratory analysis makes no reference to prior knowledge of data structure and provides models whose characteristics are determined by the statistical properties of data only.
Wavelet-based statistical analysis falls somewhat in between confirmatory and exploratory analysis since the key features of signals are discovered and analyzed with resolutions matched to their scales and subsequently assessed for statistical significance. Wavelet analysis is optimal in terms of detecting transient events and adapts well to local or nonstationary features in data within scales of the decomposition. Wavelet-based methods provide an overall richer characterization of distributed brain activation than currently employed methods.
|
|
|
|