Functional Connectivity during Resting-State Functional MR Imaging: Study of the Correspondence between Independent Component Analysis and Region-of-Interest−Based Methods

Rosazza, Cristina; Minati, Ludovico; Ghielmetti, F.; Mandelli, Maria Luisa; Bruzzone, Maria Grazia

doi:10.3174/ajnr.a2733

Cited by 106 publications

(94 citation statements)

References 35 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We considered ICA more suitable for our research question than seed-based ROI connectivity as we were interested in intrinsic networks that can be identified in a model-free, data-driven way. Most importantly, Rosazza et al (2011) showed that the largest differences in results between ICA and ROI analysis were observed for long range connections, which are the focus of our study. Moreover, ICA tends to separate signal of no interest from signal of interest (brain activation) and may give a better representation of brain activation than the raw MRI signal and contain less noise (Van de Ven et al, 2004;Fox and Raichle, 2007).…”

Section: Introductionmentioning

confidence: 76%

“…Stability of the components, i.e. investigating whether a component has the tendency to split or merge with another component (Rosazza et al, 2011), was validated by running the ICASSO toolbox implemented in GIFT using twenty iterations with both random iterations and bootstrapping (Himberg and Hyvärinen, 2003;Himberg et al, 2004).…”

Section: Discussionmentioning

confidence: 99%

“…Moreover, ICA tends to separate signal of no interest from signal of interest (brain activation) and may give a better representation of brain activation than the raw MRI signal and contain less noise (Van de Ven et al, 2004;Fox and Raichle, 2007). Finally, it may be difficult with ROI analysis to clearly define a complete brain network, without missing certain brain areas or including areas that show different temporal patterns (Fox and Raichle, 2007;Rosazza et al, 2011). Therefore, ICA is the optimal choice to study alterations in concerted action between spatially distributed brain networks (Shenton et al, 2001;Kubicki et al, 2002;Hubl et al, 2004) as we described earlier.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Abnormal connectivity between attentional, language and auditory networks in schizophrenia

Liemburg

Vercammen

Horst

et al. 2012

Schizophrenia Research

View full text Add to dashboard Cite

Section: Introductionmentioning

confidence: 76%

Section: Discussionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Abnormal connectivity between attentional, language and auditory networks in schizophrenia

Liemburg

Vercammen

Horst

et al. 2012

Schizophrenia Research

View full text Add to dashboard Cite

“…Unlike ROI analyses which rely on prior anatomic hypotheses, ICA is a data‐driven analysis. There are strengths and weaknesses inherent in both methods (Rosazza et al., 2012); therefore, future studies may wish to include network homogeneity and seed‐based functional connectivity analyses as alternate methods of examining network connectivity. In addition, we employed a 4‐minute resting state scan, which is sufficient to reliably estimate network connectivity.…”

Section: Discussionmentioning

confidence: 99%

“…We isolated 20 ICs using the Infomax algorithm in GIFT, as this number of components shows the strongest correspondence with region‐of‐interest (ROI) data (Rosazza, Minati, Ghielmetti., Mandelli, & Bruzzone, 2012), and repeated the estimation 20 times using the ICASSO method (Himberg, Hyvärinen, & Esposito, 2004) to enhance component reliability. Single subject spatial maps and time courses for each component were back‐reconstructed for each participant and converted to z ‐scores to indicate the strength of each voxel's contribution to the component (Erhardt et al., 2011).…”

Section: Methodsmentioning

confidence: 99%

Relation between patterns of intrinsic network connectivity, cognitive functioning, and symptom presentation in trauma‐exposed patients with major depressive disorder

et al. 2017

View full text Add to dashboard Cite

ObjectiveThe present study investigated resting fMRI connectivity within the default mode (DMN), salience (SN), and central executive (CEN) networks in relation to neurocognitive performance and symptom severity in trauma‐exposed patients with major depressive disorder (MDD).MethodGroup independent component analysis was conducted among patients with MDD (n = 21), examining DMN, SN, and CEN connectivity in relation to neurocognitive performance and symptom severity. Activation in these networks was also compared between the patient group and healthy controls (n = 20).ResultsAmong the patient group, higher levels of performance on measures of verbal memory and executive functioning were related to increased connectivity within the DMN (i.e., inferior parietal lobe; precuneus). Greater depression severity was related to reduced connectivity between the SN and a node of the DMN (i.e., posterior cingulate cortex) and higher depersonalization symptoms were related to enhanced connectivity between the SN and a node of the DMN (i.e., middle temporal gyrus). Higher symptoms of depersonalization were also associated with reduced integration of the DMN with the medial frontal gyrus. Relative to controls, patients with MDD showed greater connectivity of the ventromedial prefrontal cortex within the DMN.ConclusionIntrinsic connectivity network patterns are related to cognitive performance and symptom presentation among trauma‐exposed patients with MDD.

show abstract

Multimodal study of default‐mode network integrity in disorders of consciousness

Rosazza¹,

Andronache²,

Sattin

et al. 2016

Annals of Neurology

View full text Add to dashboard Cite

Objective: Understanding residual brain function in disorders of consciousness poses extraordinary challenges, and imaging examinations are needed to complement clinical assessment. The default-mode network (DMN) is known to be dysfunctional, although correlation with level of consciousness remains controversial. We investigated DMN activity with resting-state functional magnetic resonance imaging (rs-fMRI), alongside its structural and metabolic integrity, aiming to elucidate the corresponding associations with clinical assessment. Methods: We enrolled 119 consecutive patients: 72 in a vegetative state/unresponsive wakefulness state (VS/UWS), 36 in a minimally conscious state (MCS), and 11 with severe disability. All underwent structural MRI and rs-fMRI, and a subset also underwent 18 F-fluorodeoxyglucose positron emission tomography (FDG-PET). Data were analyzed with manual and automatic approaches, in relation to diagnosis and clinical score. Results: Excluding the quartile with largest head movement, DMN activity was decreased in VS/UWS compared to MCS, and correlated with clinical score. Independent-component and seed-based analyses provided similar results, although the latter and their combination were most informative. Structural MRI and FDG-PET were less sensitive to head movement and had better diagnostic accuracy than rs-fMRI only when all cases were included. rs-fMRI indicated relatively preserved DMN activity in a small subset of VS/UWS patients, 2 of whom evolved to MCS. The integrity of the left hemisphere appears to be predictive of a better clinical status. Interpretation: rs-fMRI of the DMN is sensitive to clinical severity. The effect is consistent across data analysis approaches, but heavily dependent on head movement. rs-fMRI could be informative in detecting residual DMN activity for those patients who remain relatively still during scanning and whose diagnosis is uncertain.

show abstract

Functional Connectivity during Resting-State Functional MR Imaging: Study of the Correspondence between Independent Component Analysis and Region-of-Interest−Based Methods

Cited by 106 publications

References 35 publications

Abnormal connectivity between attentional, language and auditory networks in schizophrenia

Abnormal connectivity between attentional, language and auditory networks in schizophrenia

Relation between patterns of intrinsic network connectivity, cognitive functioning, and symptom presentation in trauma‐exposed patients with major depressive disorder

Multimodal study of default‐mode network integrity in disorders of consciousness

Contact Info

Product

Resources

About