In the recent years, the field of quality data assessment and signal denoising in functional magnetic resonance imaging (fMRI) is rapidly evolving and the identification and reduction of spurious signal with pre-processing pipeline is one of the most discussed topic. In particular, subject motion or physiological signals, such as respiratory or/and cardiac pulsatility, were showed to introduce false-positive activations in subsequent statistical analyses. Different measures for the evaluation of the impact of motion related artefacts, such as frame-wise displacement and root mean square of movement parameters, and the reduction of these artefacts with different approaches, such as linear regression of nuisance signals and scrubbing or censoring procedure, were introduced. However, we identify two main drawbacks: i) the different measures used for the evaluation of motion artefacts were based on user-dependent thresholds, and ii) each study described and applied their own pre-processing pipeline. Few studies analysed the effect of these different pipelines on subsequent analyses methods in task-based fMRI.The first aim of the study is to obtain a tool for motion fMRI data assessment, based on auto-calibrated procedures, to detect outlier subjects and outliers volumes, targeted on each investigated sample to ensure homogeneity of data for motion. The second aim is to compare the impact of different pre-processing pipelines on task-based fMRI using GLM based on recent advances in resting state fMRI preprocessing pipelines. Different output measures based on signal variability and task strength were used for the assessment.
Quality data assessment and improvement in pre-processing pipeline to minimize impact of spurious signals in functional magnetic imaging (fMRI) / Nigri, Anna. - (2017).
|Titolo:||Quality data assessment and improvement in pre-processing pipeline to minimize impact of spurious signals in functional magnetic imaging (fMRI)|
|Data di pubblicazione:||2017|
|Appare nelle tipologie:||8.1 Doctoral thesis Polito|