We propose a sparse representation approach for classifying different targets in Synthetic Aperture Radar (SAR) images. Unlike the other feature based approaches, the proposed method does not require explicit pose estimation or any preprocessing. The dictionary used in this setup is the collection of the normalized training vectors itself. Computing a sparse representation for the test data using this dictionary corresponds to finding a locally linear approximation with respect to the underlying class manifold. SAR images obtained from the Moving and Stationary Target Acquisition and Recognition (MSTAR) public database were used in the classification setup. Results show that the performance of the algorithm is superior to using a support vector machines based approach with similar assumptions. Significant complexity reduction is obtained by reducing the dimensions of the data using random projections for only a small loss in performance.

Digital health data are multimodal and high-dimensional. A patient’s health state can be characterized by a multitude of signals including medical imaging, clinical variables, genome sequencing, conversations between clinicians and patients, and continuous signals from wearables, among others. This high volume, personalized data stream aggregated over patients’ lives has spurred interest in developing new artificial intelligence (AI) models for higher-precision diagnosis, prognosis, and tracking. While the promise of these algorithms is undeniable, their dissemination and adoption have been slow, owing partially to unpredictable AI model performance once deployed in the real world. We posit that one of the rate-limiting factors in developing algorithms that generalize to real-world scenarios is the very attribute that makes the data exciting—their high-dimensional nature. This paper considers how the large number of features in vast digital health data can challenge the development of robust AI models—a phenomenon known as “the curse of dimensionality” in statistical learning theory. We provide an overview of the curse of dimensionality in the context of digital health, demonstrate how it can negatively impact out-of-sample performance, and highlight important considerations for researchers and algorithm designers.

Introduction Although migraine and persistent post-traumatic headache often share phenotypic characteristics, few studies have interrogated the pathophysiological differences underlying these headache types. While there is now some indication of differences in brain structure between migraine and persistent post-traumatic headache, differences in brain function have not been adequately investigated. The objective of this study was to compare static and dynamic functional connectivity patterns in migraine versus persistent post-traumatic headache using resting-state magnetic resonance imaging. Methods This case-control study interrogated the static functional connectivity and dynamic functional connectivity patterns of 59 a priori selected regions of interest involved in pain processing. Pairwise connectivity (region of interest to region of interest) differences between migraine (n = 33) and persistent post-traumatic headache (n = 44) were determined and compared to healthy controls (n = 36) with ANOVA and subsequent t-tests. Pearson partial correlations were used to explore the relationship between headache burden (headache frequency; years lived with headache) and functional connectivity and between pain intensity at the time of imaging and functional connectivity for migraine and persistent post-traumatic headache groups, separately. Results Significant differences in static functional connectivity between migraine and persistent post-traumatic headache were found for 17 region pairs that included the following regions of interest: Primary somatosensory, secondary somatosensory, posterior insula, hypothalamus, anterior cingulate, middle cingulate, temporal pole, supramarginal gyrus, superior parietal, middle occipital, lingual gyrus, pulvinar, precuneus, cuneus, somatomotor, ventromedial prefrontal cortex, and dorsolateral prefrontal cortex. Significant differences in dynamic functional connectivity between migraine and persistent post-traumatic headache were found for 10 region pairs that included the following regions of interest: Secondary somatosensory, hypothalamus, middle cingulate, temporal pole, supramarginal gyrus, superior parietal, lingual gyrus, somatomotor, precentral, posterior cingulate, middle frontal, fusiform gyrus, parieto-occiptal, and amygdala. Although there was overlap among the regions demonstrating static functional connectivity differences and those showing dynamic functional connectivity differences between persistent post-traumatic headache and migraine, there was no overlap in the region pair functional connections. After controlling for sex and age, there were significant correlations between years lived with headache with static functional connectivity of the right dorsolateral prefrontal cortex with the right ventromedial prefrontal cortex in the migraine group and with static functional connectivity of right primary somatosensory with left supramarginal gyrus in the persistent post-traumatic headache group. There were significant correlations between headache frequency with static functional connectivity of left secondary somatosensory with right cuneus in the migraine group and with static functional connectivity of left middle cingulate with right pulvinar and right posterior insula with left hypothalamus in the persistent post-traumatic headache group. Dynamic functional connectivity was significantly correlated with headache frequency, after controlling for sex and age, in the persistent post-traumatic headache group for one region pair (right middle cingulate with right supramarginal gyrus). Dynamic functional connectivity was correlated with pain intensity at the time of imaging for the migraine cohort for one region pair (right posterior cingulate with right amygdala). Conclusions Resting-state functional imaging revealed static functional connectivity and dynamic functional connectivity differences between migraine and persistent post-traumatic headache for regions involved in pain processing. These differences in functional connectivity might be indicative of distinctive pathophysiology associated with migraine versus persistent post-traumatic headache.

BackgroundThe majority of individuals with post-traumatic headache have symptoms that are indistinguishable from migraine. The overlap in symptoms amongst these individuals raises the question as to whether post-traumatic headache has a unique pathophysiology or if head trauma triggers migraine. The objective of this study was to compare brain structure in individuals with persistent post-traumatic headache (i.e. headache lasting at least 3 months following a traumatic brain injury) attributed to mild traumatic brain injury to that of individuals with migraine.MethodsTwenty-eight individuals with persistent post-traumatic headache attributed to mild traumatic brain injury and 28 individuals with migraine underwent brain magnetic resonance imaging on a 3 T scanner. Regional volumes, cortical thickness, surface area and curvature measurements were calculated from T1-weighted sequences and compared between subject groups using ANCOVA. MRI data from 28 healthy control subjects were used to interpret the differences in brain structure between migraine and persistent post-traumatic headache.ResultsDifferences in regional volumes, cortical thickness, surface area and brain curvature were identified when comparing the group of individuals with persistent post-traumatic headache to the group with migraine. Structure was different between groups for regions within the right lateral orbitofrontal lobe, left caudal middle frontal lobe, left superior frontal lobe, left precuneus and right supramarginal gyrus (p < .05). Considering these regions only, there were differences between individuals with persistent post-traumatic headache and healthy controls within the right lateral orbitofrontal lobe, right supramarginal gyrus, and left superior frontal lobe and no differences when comparing the migraine cohort to healthy controls.ConclusionsIn conclusion, persistent post-traumatic headache and migraine are associated with differences in brain structure, perhaps suggesting differences in their underlying pathophysiology. Additional studies are needed to further delineate similarities and differences in brain structure and function that are associated with post-traumatic headache and migraine and to determine their specificity for each of the headache types.

Information divergence functions play a critical role in statistics and information theory. In this paper we show that a non-parametric f-divergence measure can be used to provide improved bounds on the minimum binary classification probability of error for the case when the training and test data are drawn from the same distribution and for the case where there exists some mismatch between training and test distributions. We confirm the theoretical results by designing feature selection algorithms using the criteria from these bounds and by evaluating the algorithms on a series of pathological speech classification tasks.

Changes in some lexical features of language have been associated with the onset and progression of Alzheimer's disease. Here we describe a method to extract key features from discourse transcripts, which we evaluated on non-scripted news conferences from President Ronald Reagan, who was diagnosed with Alzheimer's disease in 1994, and President George Herbert Walker Bush, who has no known diagnosis of Alzheimer's disease. Key word counts previously associated with cognitive decline in Alzheimer's disease were extracted and regression analyses were conducted. President Reagan showed a significant reduction in the number of unique words over time and a significant increase in conversational fillers and non-specific nouns over time. There was no significant trend in these features for President Bush.

Objective: To determine the potential for improving amyotrophic lateral sclerosis (ALS) clinical trials by having patients or caregivers perform frequent selfassessments at home. Methods and Participants: We enrolled ALS patients into a nonblinded, longitudinal 9-month study in which patients and caregivers obtained daily data using several different instruments, including a slow-vital capacity device, a hand grip dynamometer, an electrical impedance myographybased fitness device, an activity tracker, a speech app, and the ALS functional rating scale-revised. Questions as to acceptability were asked at two time points. Results: A total of 113 individuals enrolled, with 61 (43 men, 18 women, mean age 60.1 AE 9.9 years) collecting a minimum of 7 days data and being included in the analysis. Daily measurements resulted in more accurate assessments of the slope of progression of the disease, resulting in smaller sample size estimates for a hypothetical clinical trial. For example, by performing daily slow-vital capacity measurements, calculated sample size was reduced to 182 subjects/ study arm from 882/arm for monthly measurements. Similarly, performing the ALS functional rating scale weekly rather than monthly led to a calculated sample size of 73/arm as compared to 274/arm. Participants generally found the procedures acceptable and, for many, improved their sense of control of their disease. Interpretation: Frequent at-home measurements using standard tools holds the prospect of tracking progression and reducing sample size requirements for clinical trials in ALS while also being acceptable to the patients. Future studies in this and other neurological disorders should consider adopting this approach to data collection.

Objective: To design an ALS clinical study in which patients are remotely recruited, screened, enrolled and then assessed via daily data collection at home by themselves or caregivers. Methods: This observational, natural-history study included two academic medical centers, one providing overall clinical management and the other overseeing computing and webservices design and management. Both healthy and ALS subjects were recruited on the Internet via advertisement on governmental and foundation websites as well as through Facebook and Google paid advertisements. Individuals underwent screening and enrollment remotely, including signing an electronic informed consent form. Participants were then provided self-measurement equipment and instructed on their use through a series of web-based videos. The equipment included a handgrip dynamometer, spirometer with smartphone connection, electrical impedance myography device, and an activity tracker. ALS Functional Rating Scale-Revised data were also collected. Subjects were asked to collect data daily for three months and twice-weekly for the subsequent six months. Results: One hundred and eleven ALS patients and 30 healthy individuals enrolled in the study from across 41 states (74 men, 62 women). Baseline median ALSFRS-R score was 33. Seventy two percent of the ALS patients sent equipment and 88% of the healthy subjects sent equipment were able to complete a first set of measurements. Expected baseline differences between the ALS patients and healthy participants were identified for all measures. Conclusions: It is possible to design and institute an at-home based study in ALS patients, using a number of state-of-the-art approaches, including web-based consenting and training and Internet-connected measurement devices.