Classification of Alzheimer's Progression Using fMRI Data.

In the last three decades, the development of functional magnetic resonance imaging (fMRI) has significantly contributed to the understanding of the brain, functional brain mapping, and resting-state brain networks. Given the recent successes of deep learning in various fields, we propose a 3D-CNN-LSTM classification model to diagnose health conditions with the following classes: condition normal (CN), early mild cognitive impairment (EMCI), late mild cognitive impairment (LMCI), and Alzheimer's disease (AD). The proposed method employs spatial and temporal feature extractors, wherein the former utilizes a U-Net architecture to extract spatial features, and the latter utilizes long short-term memory (LSTM) to extract temporal features. Prior to feature extraction, we performed four-step pre-processing to remove noise from the fMRI data. In the comparative experiments, we trained each of the three models by adjusting the time dimension. The network exhibited an average accuracy of 96.4% when using five-fold cross-validation. These results show that the proposed method has high potential for identifying the progression of Alzheimer's by analyzing 4D fMRI data.

Reward biases spontaneous neural reactivation during sleep.

Sleep favors the reactivation and consolidation of newly acquired memories. Yet, how our brain selects the noteworthy information to be reprocessed during sleep remains largely unknown. From an evolutionary perspective, individuals must retain information that promotes survival, such as avoiding dangers, finding food, or obtaining praise or money. Here, we test whether neural representations of rewarded (compared to non-rewarded) events have priority for reactivation during sleep. Using functional MRI and a brain decoding approach, we show that patterns of brain activity observed during waking behavior spontaneously reemerge during slow-wave sleep. Critically, we report a privileged reactivation of neural patterns previously associated with a rewarded task (i.e., winning at a complex game). Moreover, during sleep, activity in task-related brain regions correlates with better subsequent memory performance. Our study uncovers a neural mechanism whereby rewarded life experiences are preferentially replayed and consolidated while we sleep.

Sign language recognition with the Kinect sensor based on conditional random fields.

Sign language is a visual language used by deaf people. One difficulty of sign language recognition is that sign instances of vary in both motion and shape in three-dimensional (3D) space. In this research, we use 3D depth information from hand motions, generated from Microsoft's Kinect sensor and apply a hierarchical conditional random field (CRF) that recognizes hand signs from the hand motions. The proposed method uses a hierarchical CRF to detect candidate segments of signs using hand motions, and then a BoostMap embedding method to verify the hand shapes of the segmented signs. Experiments demonstrated that the proposed method could recognize signs from signed sentence data at a rate of 90.4%.

Sign language spotting with a threshold model based on conditional random fields.

Sign language spotting is the task of detecting and recognizing signs in a signed utterance, in a set vocabulary. The difficulty of sign language spotting is that instances of signs vary in both motion and appearance. Moreover, signs appear within a continuous gesture stream, interspersed with transitional movements between signs in a vocabulary and nonsign patterns (which include out-of-vocabulary signs, epentheses, and other movements that do not correspond to signs). In this paper, a novel method for designing threshold models in a conditional random field (CRF) model is proposed which performs an adaptive threshold for distinguishing between signs in a vocabulary and nonsign patterns. A short-sign detector, a hand appearance-based sign verification method, and a subsign reasoning method are included to further improve sign language spotting accuracy. Experiments demonstrate that our system can spot signs from continuous data with an 87.0 percent spotting rate and can recognize signs from isolated data with a 93.5 percent recognition rate versus 73.5 percent and 85.4 percent, respectively, for CRFs without a threshold model, short-sign detection, subsign reasoning, and hand appearance-based sign verification. Our system can also achieve a 15.0 percent sign error rate (SER) from continuous data and a 6.4 percent SER from isolated data versus 76.2 percent and 14.5 percent, respectively, for conventional CRFs.