Estimating repetitive spatiotemporal patterns from many subjects' resting-state fMRIs

Interpretation of deep non-linear factorization for autism

Autism, a neurodevelopmental disorder, presents significant challenges for diagnosis and classification. Despite the widespread use of neural networks in autism classification, the interpretability of their models remains a crucial issue. This study aims to address this concern by investigating the interpretability of neural networks in autism classification using the deep symbolic regression and brain network interpretative methods. Specifically, we analyze publicly available autism fMRI data using our previously developed Deep Factor Learning model on a Hibert Basis tensor (HB-DFL) method and extend the interpretative Deep Symbolic Regression method to identify dynamic features from factor matrices, construct brain networks from generated reference tensors, and facilitate the accurate diagnosis of abnormal brain network activity in autism patients by clinicians. Our experimental results show that our interpretative method effectively enhances the interpretability of neural networks and identifies crucial features for autism classification.

Moving beyond the 'CAP' of the Iceberg: Intrinsic connectivity networks in fMRI are continuously engaging and overlapping

Resting-state functional magnetic resonance imaging is currently the mainstay of functional neuroimaging and has allowed researchers to identify intrinsic connectivity networks (aka functional networks) at different spatial scales. However, little is known about the temporal profiles of these networks and whether it is best to model them as continuous phenomena in both space and time or, rather, as a set of temporally discrete events. Both categories have been supported by series of studies with promising findings. However, a critical question is whether focusing only on time points presumed to contain isolated neural events and disregarding the rest of the data is missing important information, potentially leading to misleading conclusions. In this work, we argue that brain networks identified within the spontaneous blood oxygenation level-dependent (BOLD) signal are not limited to temporally sparse burst moments and that these event present time points (EPTs) contain valuable but incomplete information about the underlying functional patterns. We focus on the default mode and show evidence that is consistent with its continuous presence in the BOLD signal, including during the event absent time points (EATs), i.e., time points that exhibit minimum activity and are the least likely to contain an event. Moreover, our findings suggest that EPTs may not contain all the available information about their corresponding networks. We observe distinct default mode connectivity patterns obtained from all time points (AllTPs), EPTs, and EATs. We show evidence of robust relationships with schizophrenia symptoms that are both common and unique to each of the sets of time points (AllTPs, EPTs, EATs), likely related to transient patterns of connectivity. Together, these findings indicate the importance of leveraging the full temporal data in functional studies, including those using event-detection approaches.

Whole-brain propagating patterns in human resting-state brain activities

Repetitive propagating activities in resting-state brain activities have been widely observed in various species and regions. Because they resemble the preceding brain activities during tasks, they are assumed to reflect past experiences embedded in neuronal circuits. "Whole-brain" propagating activities may also reflect a process that integrates information distributed over the entire brain, such as visual and motor information. Here we reveal whole-brain propagating activities from human resting-state magnetoencephalography (MEG) and electroencephalography (EEG) data. We simultaneously recorded the MEGs and EEGs and estimated the source currents from both measurements. Then using our recently proposed algorithm, we extracted repetitive spatiotemporal patterns from the source currents. The estimated patterns consisted of multiple frequency components, each of which transiently exhibited the frequency-specific resting-state networks (RSNs) of functional MRIs (fMRIs), such as the default mode and sensorimotor networks. A simulation test suggested that the spatiotemporal patterns reflected the phase alignment of the multiple frequency oscillators induced by the propagating activities along the anatomical connectivity. These results argue that whole-brain propagating activities transiently exhibited multiple RSNs in their multiple frequency components, suggesting that they reflected a process to integrate the information distributed over the frequencies and networks.