Alessio Medda

A wavelet clustering technique for the identification of functionally connected regions in the rat brain using resting state fMRI

This work presents a new data-driven method for the identification of functionally connected regions in the rat brain, using agglomerative clustering based on the discrete wavelet transform (DWT). The proposed approach is evaluated on resting state fMRI data and no a priori assumptions about the distribution of the signals or anatomical ROIs are made. The coefficients of the DWT are used as features in the clustering algorithm, and the performance of the classifier is evaluated as the capability to produce clusters that best correlate with known anatomical regions in the sensorimotor cortex of the brain. Wavelet features that best represent salient characteristics in the spectrum of the voxel signals are found to produce best clustering results.

Multiscale FC analysis refines functional connectivity networks in individual brains

Recent advances in functional connectivity (FC) analysis of functional magnetic resonance imaging (fMRI) data facilitate the characterization of the brains intrinsic functional networks (FC-fMRI). Because the fMRI signal does not provides a perfect representation of neuronal activity, the potential for FC-fMRI to identify functionally relevant networks critically depends upon separating overlapping signals from one another and from external noise. As a step in data preconditioning, researchers often band-pass filter fMRI signals to the range from 0.01 Hz to 0.1 Hz. However, coordinated network oscillations operate across multiple frequencies. Thus, it is not clear that the view of FC-fMRI networks within a single spectral range produces the fullest characterization of brains multiple and overlapping systems. The following study addresses this limitation by advancing a multiscale fractionation of FC-fMRI networks, as well methods for quantifying cross-spectral network similarity. These methods clearly and consistently represent group-level brains as composed of well-known functional networks.

The Integrated Blast Effects Sensor Suite: A rapidly developed, complex, system of systems

The need for rapid development of tactical system of systems solutions for military applications requires the use of system modeling techniques and simulation and validation methods to be applied throughout the lifecycle of the system. This combined approach of development and verification is preferred to traditional approaches for risk mitigation and cost effectiveness. This paper examines the Integrated Blast Effects Sensor Suite developed at the Georgia Tech Research Institute and its architecture as a complex system of systems.

Wavelet packet based clustering for the study of functional connectivity in the rat brain

This paper presents a data-driven clustering method based on the use of wavelet packet features for the study of functionally connected regions in the brain. In particular, wavelet packets are used because of their optimal whitening properties for 1/f-like processes, in association with a uniform segmentation of the frequency axis. Features obtained by the wavelet packet transform are grouped together using agglomerative clustering with standardize Euclidian distance. The results obtained are compared with known atlas of the rat brain and with results obtained from a technique based on the standard wavelet transform previously presented by the same authors. The presented approach produces clusters that are well matched to classical anatomical regions in the sensorimotor cortex of the rat brain and also produces clusters that roughly correspond to primary motor, primary somatosensory and secondary somatosensory areas, with the subcortical clusters typically comprising one or two groups in the caudate putamen area.

Instantaneous brain dynamics mapped to a continuous state space

&NA; Measures of whole‐brain activity, from techniques such as functional Magnetic Resonance Imaging, provide a means to observe the brains dynamical operations. However, interpretation of whole‐brain dynamics has been stymied by the inherently high‐dimensional structure of brain activity. The present research addresses this challenge through a series of scale transformations in the spectral, spatial, and relational domains. Instantaneous multispectral dynamics are first developed from input data via a wavelet filter bank. Voxel‐level signals are then projected onto a representative set of spatially independent components. The correlation distance over the instantaneous wavelet‐ICA state vectors is a graph that may be embedded onto a lower‐dimensional space to assist the interpretation of state‐space dynamics. Applying this procedure to a large sample of resting‐state and task‐active data (acquired through the Human Connectome Project), we segment the empirical state space into a continuum of stimulus‐dependent brain states. Upon observing the local neighborhood of brain‐states adopted subsequent to each stimulus, we may conclude that resting brain activity includes brain states that are, at times, similar to those adopted during tasks, but that are at other times distinct from task‐active brain states. As task‐active brain states often populate a local neighborhood, back‐projection of segments of the dynamical state space onto the brains surface reveals the patterns of brain activity that support many experimentally‐defined states. HighlightsWe demonstrate the construction and interrogation of a continuous, two‐dimensional map of fMRI dynamics.Map points represent an individuals multispectral, and multispectral BOLD state centered at a single point in time.Task‐based scans occupy focal state‐spaces, reinforcing the utility of study methods to capture salient BOLD dynamics.Resting‐state scans occupy a broad state‐space, reinforcing the view that the resting mind is highly active.

Wavelet-based clustering of resting state MRI data in the rat

While functional connectivity has typically been calculated over the entire length of the scan (5-10min), interest has been growing in dynamic analysis methods that can detect changes in connectivity on the order of cognitive processes (seconds). Previous work with sliding window correlation has shown that changes in functional connectivity can be observed on these time scales in the awake human and in anesthetized animals. This exciting advance creates a need for improved approaches to characterize dynamic functional networks in the brain. Previous studies were performed using sliding window analysis on regions of interest defined based on anatomy or obtained from traditional steady-state analysis methods. The parcellation of the brain may therefore be suboptimal, and the characteristics of the time-varying connectivity between regions are dependent upon the length of the sliding window chosen. This manuscript describes an algorithm based on wavelet decomposition that allows data-driven clustering of voxels into functional regions based on temporal and spectral properties. Previous work has shown that different networks have characteristic frequency fingerprints, and the use of wavelets ensures that both the frequency and the timing of the BOLD fluctuations are considered during the clustering process. The method was applied to resting state data acquired from anesthetized rats, and the resulting clusters agreed well with known anatomical areas. Clusters were highly reproducible across subjects. Wavelet cross-correlation values between clusters from a single scan were significantly higher than the values from randomly matched clusters that shared no temporal information, indicating that wavelet-based analysis is sensitive to the relationship between areas.

The Integrated Blast Effects Sensor Suite: A Rapidly Developed, Complex, System of Systems

Spurned by the increasing concern and consciousness of traumatic brain injuries in deployed U.S. service members, the U.S. Army Rapid Equipping Force sought help from the Georgia Tech Research Institute to rapidly develop and deploy a system capable of gathering relevant soldier-centric data-the Integrated Blast Effects Sensor Suite. To meet aggressive program milestones and requirements, Georgia Tech Research Institute engaged in rapid systems engineering efforts focused on leveraging iterative development and test methodologies. Ultimately, an integrated system of systems composed of vehicle systems, soldier-worn headset and torso systems, and data retrieval systems was deployed to troops in Afghanistan for an operational assessment. The Integrated Blast Effects Sensor Suite development process and parallel efforts investigating injury dosimetry methodologies have yielded unique findings and lessons learned, which should be incorporated into future evolutions of similar systems.

Multiscale functional networks in human resting state functional MRI

Recent advent of fast imaging techniques for MRI application allow whole brain coverage with sub-second resolution, opening the door for new data-driven computational techniques that can harvest the information contained in the data. This paper examines the use of wavelet based spectral decomposition and hierarchical clustering for resting state functional MRI. Wavelet packets naturally enable short time spectral decomposition with minimal temporal window lengths across multiple frequency ranges, while hierarchical clustering is used for organizing broadband and filtered fMRI data into functional network. This method was applied to human group data from five volunteers from the 1000 Functional Connectomes database.

Agglomerative clustering for resting state MRI

Methods to interpret data obtained from resting state functional magnetic imaging (rs-fMRI) must be developed to more thoroughly understand how network structure of the brain supports the body and the mind. To this end, we examine the use of agglomerative clustering (AC) as a method for rs-fMRI analysis. AC is a data driven approach for organizing spatially distinct clusters of temporally similar activity. Its application to rs-fMRI data produces spatial parcellation of brain areas that share similar temporal characteristics. The technique is scalable, enabling identification of local to widespread organization. Using a wavelet based filter bank, the technique is made amenable to frequency domain scaling as well. Comparisons drawn between AC and two alternative rs-fMRI analytics - seed-based correlation, and spatial independent component analysis - highlight the ability of the proposed technique to recognize well known functional brain regions.

Virtually transforming detect: moving a 2-D gamified health assessment to virtual reality

Researchers from Georgia Tech, Emory University, and Georgia Tech Research Institute previously collaborated on a 2D Heads-Up Display based gamified health assessment focused on determining neurologic impairment. As a part of continuing gamification of the assessment, the entire project was translated to 3D Virtual Reality under significant constraints. The system needed to test the same neural pathways in the same manner as the previous assessment in order to use data from the 2D system as validation for the approach. Some aspects of user response to immersion required significant changes to gameplay mechanics. Iterative, collaborative development was critical to increasing gamification while maintaining an effective assessment tool.