Leigh A. Johnston

Performance analysis of a dynamic programming track before detect algorithm

We analyze a dynamic programming (DP)-based track before detect (TBD) algorithm. By using extreme value theory we obtain explicit expressions for various performance measures of the algorithm such as probability of detection and false alarm. Our analysis has two advantages. First the unrealistic Gaussian and independence assumptions used in previous works are not required. Second, the probability of detection and false alarm curves obtained fit computer simulated performance results significantly more accurately than previously proposed analyses of the TBD algorithm.

An improvement to the interacting multiple model (IMM) algorithm

Computing the optimal conditional mean state estimate for a jump Markov linear system requires exponential complexity, and hence, practical filtering algorithms are necessarily suboptimal. In the target tracking literature, suboptimal multiple-model filtering algorithms, such as the interacting multiple model (IMM) method and generalized pseudo-Bayesian (GPB) schemes, are widely used for state estimation of such systems. We derive a reweighted interacting multiple model algorithm. Although the IMM algorithm is an approximation of the conditional mean state estimator, our algorithm is a recursive implementation of a maximum a posteriori (MAP) state sequence estimator. This MAP estimator is an instance of a previous version of the EM algorithm known as the alternating expectation conditional maximization (AECM) algorithm. Computer simulations indicate that the proposed reweighted IMM algorithm is a competitive alternative to the popular IMM algorithm and GPB methods.

The Power of Spectral Density Analysis for Mapping Endogenous BOLD Signal Fluctuations

FMRI has revealed the presence of correlated low‐frequency cerebro‐vascular oscillations within functional brain systems, which are thought to reflect an intrinsic feature of large‐scale neural activity. The spatial correlations shown by these fluctuations has been their identifying feature, distinguishing them from fluctuations associated with other processes. Major analysis methods characterize these correlations, identifying networks and their interactions with various factors. However, other analysis approaches are required to fully characterize the regional signal dynamics contributing to these correlations between regions. In this study we show that analysis of the power spectral density (PSD) of regional signals can identify changes in oscillatory dynamics across conditions, and is able to characterize the nature and spatial extent of signal changes underlying changes in measures of connectivity. We analyzed spectral density changes in sessions consisting of both resting‐state scans and scans recording 2 min blocks of continuous unilateral finger tapping and rest. We assessed the relationship of PSD and connectivity measures by additionally tracking correlations between selected motor regions. Spectral density gradually increased in gray and white matter during the experiment. Finger tapping produced widespread decreases in low‐frequency spectral density. This change was symmetric across the cortex, and extended beyond both the lateralized task‐related signal increases, and the established “resting‐state” motor network. Correlations between motor regions also reduced with task performance. In conclusion, analysis of PSD is a sensitive method for detecting and characterizing BOLD signal oscillations that can enhance the analysis of network connectivity. Hum Brain Mapp 2008.

Contour grouping with prior models

Conventional approaches to perceptual grouping assume little specific knowledge about the object(s) of interest. However, there are many applications in which such knowledge is available and useful. Here, we address the problem of finding the bounding contour of an object in an image when some prior knowledge about the object is available. We introduce a framework for combining prior probabilistic knowledge of the appearance of the object with probabilistic models for contour grouping. A constructive search technique is used to compute candidate closed object boundaries, which are then evaluated by combining figure, ground, and prior probabilities to compute the maximum a posteriori estimate. A significant advantage of our formulation is that it rigorously combines probabilistic local cues with important global constraints such as simplicity (no self-intersections), closure, completeness, and nontrivial scale priors. We apply this approach to the problem of computing exact lake boundaries from satellite imagery, given approximate prior knowledge from an existing digital database. We quantitatively evaluate the performance of our algorithm and find that it exceeds the performance of human mapping experts and a competing active contour approach, even with relatively weak prior knowledge. While the priors may be task-specific, the approach is general, as we demonstrate by applying it to a completely different problem: the computation of human skin boundaries in natural imagery.

Opportunistic file transfer over a fading channel: A POMDP search theory formulation with optimal threshold policies

We present a computationally efficient algorithm that minimizes the transmission energy and latency associated with transmitting a file across a Gilbert Elliott fading channel. We formulate the optimal tradeoff between transmission energy and latency as a partially observed Markov decision process problem (POMDP). The channel state is not directly observed and hence transmission decisions must be based on ACK/NAK information provided over a feedback channel. The key idea is to reformulate the resulting POMDP as a Markovian search problem, with optimal transmission control policies that are threshold in nature. Threshold policies are computationally inexpensive to implement. Our analysis shows that for different parameter values of the Gilbert Elliott fading channel, the optimal transmission policy, while threshold in structure, exhibits vastly different behaviour - from persistent retransmission to back-off and wait. Numerical examples demonstrate the performance improvements that can be obtained using the optimal threshold policies as compared to existing heuristic algorithms.

Diffusion Tensor Imaging in Huntington’s disease reveals distinct patterns of white matter degeneration associated with motor and cognitive deficits

White matter (WM) degeneration is an important feature of Huntington’s disease (HD) neuropathology. To investigate WM degeneration we used Diffusion Tensor Imaging and Tract-Based Spatial Statistics to compare Fractional Anisotropy, Mean Diffusivity (MD), parallel diffusivity and perpendicular diffusivity (λ⊥) in WM throughout the whole brain in 17 clinically diagnosed HD patients and 16 matched controls. Significant WM diffusivity abnormalities were identified primarily in the corpus callosum (CC) and external/extreme capsules in HD patients compared to controls. Significant correlations were observed between motor symptoms and MD in the CC body, and between global cognitive impairment and λ⊥ in the CC genu. Probabilistic tractography from these regions revealed degeneration of functionally relevant interhemispheric WM tracts. Our findings suggest that WM degeneration within interhemispheric pathways plays an important role in the deterioration of cognitive and motor function in HD patients, and that improved understanding of WM pathology early in the disease is required.

A longitudinal diffusion tensor imaging study in symptomatic Huntington's disease

Objective The striatum and its projections are thought to be the earliest sites of Huntingtons disease (HD) pathology. This study aimed to investigate progression of striatal pathology in symptomatic HD using diffusion tensor imaging. Method Diffusion weighted images were acquired in 18 HD patients and in 17 healthy controls twice, 1 year apart. Mean diffusivity (MD) was calculated in the caudate, putamen, thalamus and corpus callosum, and compared between groups. In addition, caudate width was measured using T1 high resolution images and correlated with caudate MD. Correlation analyses were also performed in HD between caudate/putamen MD and clinical measures. Results MD was significantly higher in the caudate and putamen bilaterally for patients compared with controls at both time points although there were no significant MD differences in the thalamus or corpus callosum. For both groups, MD did not change significantly in any region from baseline to year 1. There was a significant negative correlation between caudate width and MD in patients at baseline but no correlation between these parameters in controls. There was also a significant negative correlation between Mini-Mental State Examination scores and caudate MD and putamen MD at both time points in HD. Conclusions It appears that microstructural changes influence cognitive status in HD. Although MD was significantly higher in HD compared with controls at both time points, there were no longitudinal changes in either group. This finding does not rule out the possibility that MD could be a sensitive biomarker for detecting early change in preclinical HD.

Sodium selenate reduces hyperphosphorylated tau and improves outcomes after traumatic brain injury

Traumatic brain injury is a common and serious neurodegenerative condition that lacks a pharmaceutical intervention to improve long-term outcome. Hyperphosphorylated tau is implicated in some of the consequences of traumatic brain injury and is a potential pharmacological target. Protein phosphatase 2A is a heterotrimeric protein that regulates key signalling pathways, and protein phosphatase 2A heterotrimers consisting of the PR55 B-subunit represent the major tau phosphatase in the brain. Here we investigated whether traumatic brain injury in rats and humans would induce changes in protein phosphatase 2A and phosphorylated tau, and whether treatment with sodium selenate-a potent PR55 activator-would reduce phosphorylated tau and improve traumatic brain injury outcomes in rats. Ninety young adult male Long-Evans rats were administered either a fluid percussion injury or sham-injury. A proportion of rats were killed at 2, 24, and 72 h post-injury to assess acute changes in protein phosphatase 2A and tau. Other rats were given either sodium selenate or saline-vehicle treatment that was continuously administered via subcutaneous osmotic pump for 12 weeks. Serial magnetic resonance imaging was acquired prior to, and at 1, 4, and 12 weeks post-injury to assess evolving structural brain damage and axonal injury. Behavioural impairments were assessed at 12 weeks post-injury. The results showed that traumatic brain injury in rats acutely reduced PR55 expression and protein phosphatase 2A activity, and increased the expression of phosphorylated tau and the ratio of phosphorylated tau to total tau. Similar findings were seen in post-mortem brain samples from acute human traumatic brain injury patients, although many did not reach statistical significance. Continuous sodium selenate treatment for 12 weeks after sham or fluid percussion injury in rats increased protein phosphatase 2A activity and PR55 expression, and reduced the ratio of phosphorylated tau to total tau, attenuated brain damage, and improved behavioural outcomes in rats given a fluid percussion injury. Notably, total tau levels were decreased in rats 12 weeks after fluid percussion injury, and several other factors, including the use of anaesthetic, the length of recovery time, and that some brain injury and behavioural dysfunction still occurred in rats treated with sodium selenate must be considered in the interpretation of this study. However, taken together these data suggest protein phosphatase 2A and hyperphosphorylated tau may be involved in the neurodegenerative cascade of traumatic brain injury, and support the potential use of sodium selenate as a novel traumatic brain injury therapy.

Automated differentiation of pre-diagnosis Huntington's disease from healthy control individuals based on quadratic discriminant analysis of the basal ganglia: the IMAGE-HD study

We investigated two measures of neural integrity, T1-weighted volumetric measures and diffusion tensor imaging (DTI), and explored their combined potential to differentiate pre-diagnosis Huntingtons disease (pre-HD) individuals from healthy controls. We applied quadratic discriminant analysis (QDA) to discriminate pre-HD individuals from controls and we utilised feature selection and dimension reduction to increase the robustness of the discrimination method. Thirty six symptomatic HD (symp-HD), 35 pre-HD, and 36 control individuals participated as part of the IMAGE-HD study and underwent T1-weighted MRI, and DTI using a Siemens 3 Tesla scanner. Volume and DTI measures [mean diffusivity (MD) and fractional anisotropy (FA)] were calculated for each group within five regions of interest (ROI; caudate, putamen, pallidum, accumbens and thalamus). QDA was then performed in a stepwise manner to differentiate pre-HD individuals from controls, based initially on unimodal analysis of motor or neurocognitive measures, or on volume, MD or FA measures from within the caudate, pallidum and putamen. We then tested for potential improvements to this model, by examining multi-modal MRI classifications (volume, FA and MD), and also included motor and neurocognitive measures, and additional brain regions (i.e., accumbens and thalamus). Volume, MD and FA differed across the three groups, with pre-HD characterised by significant volumetric reductions and increased FA within caudate, putamen and pallidum, relative to controls. The QDA results demonstrated that the differentiation of pre-HD from controls was highly accurate when both volumetric and diffusion data sets from basal ganglia (BG) regions were used. The highest discriminative accuracy however was achieved in a multi-modality approach and when including all available measures: motor and neurocognitive scores and multi-modal MRI measures from the BG, accumbens and thalamus. Our QDA findings provide evidence that combined multi-modal imaging measures can accurately classify individuals up to 15 years prior to onset when therapeutic intervention is likely to have maximal effects in slowing the trajectory of disease development.

Nonlinear estimation of the BOLD signal.

Signal variations in functional Magnetic Resonance Imaging experiments essentially reflect the vascular system response to increased demand for oxygen caused by neuronal activity, termed the blood oxygenation level dependent (BOLD) effect. The most comprehensive model to date of the BOLD signal is formulated as a mixed continuous-discrete-time system of nonlinear stochastic differential equations. Previous approaches to the analysis of this system have been based on linearised approximations of the dynamics, which are limited in their ability to capture the inherent nonlinearities in the physiological system. In this paper we present a nonlinear filtering method for simultaneous estimation of the hidden physiological states and the system parameters, based on an iterative coordinate descent framework. State estimates of the cerebral blood flow, cerebral blood volume and deoxyhaemoglobin content are determined using a particle filter, demonstrated via simulation to be accurate, robust and efficient in comparison to linearisation-based techniques. The adaptive state and parameter estimation algorithm generates physiologically reasonable parameter estimates for experimental fMRI data. It is anticipated that signal processing techniques for modelling and estimation will become increasingly important in fMRI analyses as limitations of linear and linearised modelling are reached.