Neuronal disorders could result in changes at brain connectomics level. One of the main goal of structural connectomics analysis deals with the possibility to detect the changes at neuronal level by studying brain connectivity. These result is possible by generating the tractography maps from diffusion MRI data, however, the algorithms for this step are affected by high sensitivity and low specificity. As a consequence, this technique could lead to result not completely reliable, with many false positive. New techniques based on tractograms filtering have been developed to resolve the problem. This thesis studies the impact of tractogram filtering for rats brain diffu- sion data (divided in rats with and without hearing problems) analyzing the changes on structural connectivity derived from tractograms considering the hearing cortex areas. In particular, Spherical-deconvolution informed filtering of tractograms (SIFT) has been used as filter. Furthermore, after the analysis of structural connectivity, a statistical based approach has been uses to classify the connectivity matrices to analyse if there was any difference between unfiltered and filtered data performance. Metrics extracted from the connectivity matrices showed differences between the unfiltered and filtered data considering the single nodes. This means the connectomes changed and, in particular, the main connections. Connectomes weighted by Fractional Anisotropy (FA), Mean Diffusivity (MD) and Radial Diffusivity (RD) showed similar results. Statistical analysis based on classifica- tion performed better in average with filtered matrices as input. The biggest difference has been observed with raw data, 27.27% more accurate for filtered data. These are the limitations of the study: number of subjects was small, Region of Interest (ROI) masks were hand made (not default atlas) and needed manual registration.

Neuronal disorders could result in changes at brain connectomics level. One of the main goal of structural connectomics analysis deals with the possibility to detect the changes at neuronal level by studying brain connectivity. These result is possible by generating the tractography maps from diffusion MRI data, however, the algorithms for this step are affected by high sensitivity and low specificity. As a consequence, this technique could lead to result not completely reliable, with many false positive. New techniques based on tractograms filtering have been developed to resolve the problem. This thesis studies the impact of tractogram filtering for rats brain diffu- sion data (divided in rats with and without hearing problems) analyzing the changes on structural connectivity derived from tractograms considering the hearing cortex areas. In particular, Spherical-deconvolution informed filtering of tractograms (SIFT) has been used as filter. Furthermore, after the analysis of structural connectivity, a statistical based approach has been uses to classify the connectivity matrices to analyse if there was any difference between unfiltered and filtered data performance. Metrics extracted from the connectivity matrices showed differences between the unfiltered and filtered data considering the single nodes. This means the connectomes changed and, in particular, the main connections. Connectomes weighted by Fractional Anisotropy (FA), Mean Diffusivity (MD) and Radial Diffusivity (RD) showed similar results. Statistical analysis based on classifica- tion performed better in average with filtered matrices as input. The biggest difference has been observed with raw data, 27.27% more accurate for filtered data. These are the limitations of the study: number of subjects was small, Region of Interest (ROI) masks were hand made (not default atlas) and needed manual registration.

Influence of tractogram filtering in analysis of tractography data in rat brains

MASCHIO, FILIPPO
2021/2022

Abstract

Neuronal disorders could result in changes at brain connectomics level. One of the main goal of structural connectomics analysis deals with the possibility to detect the changes at neuronal level by studying brain connectivity. These result is possible by generating the tractography maps from diffusion MRI data, however, the algorithms for this step are affected by high sensitivity and low specificity. As a consequence, this technique could lead to result not completely reliable, with many false positive. New techniques based on tractograms filtering have been developed to resolve the problem. This thesis studies the impact of tractogram filtering for rats brain diffu- sion data (divided in rats with and without hearing problems) analyzing the changes on structural connectivity derived from tractograms considering the hearing cortex areas. In particular, Spherical-deconvolution informed filtering of tractograms (SIFT) has been used as filter. Furthermore, after the analysis of structural connectivity, a statistical based approach has been uses to classify the connectivity matrices to analyse if there was any difference between unfiltered and filtered data performance. Metrics extracted from the connectivity matrices showed differences between the unfiltered and filtered data considering the single nodes. This means the connectomes changed and, in particular, the main connections. Connectomes weighted by Fractional Anisotropy (FA), Mean Diffusivity (MD) and Radial Diffusivity (RD) showed similar results. Statistical analysis based on classifica- tion performed better in average with filtered matrices as input. The biggest difference has been observed with raw data, 27.27% more accurate for filtered data. These are the limitations of the study: number of subjects was small, Region of Interest (ROI) masks were hand made (not default atlas) and needed manual registration.
2021
Influence of tractogram filtering in analysis of tractography data in rat brains
Neuronal disorders could result in changes at brain connectomics level. One of the main goal of structural connectomics analysis deals with the possibility to detect the changes at neuronal level by studying brain connectivity. These result is possible by generating the tractography maps from diffusion MRI data, however, the algorithms for this step are affected by high sensitivity and low specificity. As a consequence, this technique could lead to result not completely reliable, with many false positive. New techniques based on tractograms filtering have been developed to resolve the problem. This thesis studies the impact of tractogram filtering for rats brain diffu- sion data (divided in rats with and without hearing problems) analyzing the changes on structural connectivity derived from tractograms considering the hearing cortex areas. In particular, Spherical-deconvolution informed filtering of tractograms (SIFT) has been used as filter. Furthermore, after the analysis of structural connectivity, a statistical based approach has been uses to classify the connectivity matrices to analyse if there was any difference between unfiltered and filtered data performance. Metrics extracted from the connectivity matrices showed differences between the unfiltered and filtered data considering the single nodes. This means the connectomes changed and, in particular, the main connections. Connectomes weighted by Fractional Anisotropy (FA), Mean Diffusivity (MD) and Radial Diffusivity (RD) showed similar results. Statistical analysis based on classifica- tion performed better in average with filtered matrices as input. The biggest difference has been observed with raw data, 27.27% more accurate for filtered data. These are the limitations of the study: number of subjects was small, Region of Interest (ROI) masks were hand made (not default atlas) and needed manual registration.
dMRI
tractography
brain connectivity
File in questo prodotto:
File Dimensione Formato  
Maschio_Filippo.pdf

Open Access dal 16/12/2023

Dimensione 1.06 MB
Formato Adobe PDF
1.06 MB Adobe PDF Visualizza/Apri

The text of this website © Università degli studi di Padova. Full Text are published under a non-exclusive license. Metadata are under a CC0 License

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/20.500.12608/41523