Najafi, M. | Search bioRxiv

bioRxiv2017

Deep Phenotyping: Deep Learning For Temporal Phenotype/Genotype Classification

High resolution and high throughput, genotype to phenotype studies in plants are underway to accelerate breeding of climate ready crops. Complex developmental phenotypes are observed by imaging a variety of accessions in different environment conditions, however extracting the genetically heritable traits is challenging. In the recent years, deep learning techniques and in particular Convolutional Neural Networks (CNNs), Recurrent Neural Networks (RNNs) and Long-Short Term Memories (LSTMs), have shown great success in visual data recognition, classification, and sequence learning tasks. In this paper, we proposed a CNN-LSTM framework for plant classification of various genotypes. Here, we exploit the power of deep CNNs for joint feature and classifier learning, within an automatic phenotyping scheme for genotype classification. Further, plant growth variation over time is also important in phenotyping their dynamic behavior. This was fed into the deep learning framework using LSTMs to model these temporal cues for different plant accessions. We generated a replicated dataset of four accessions of Arabidopsis and carried out automated phenotyping experiments. The results provide evidence of the benefits of our approach over using traditional hand-crafted image analysis features and other genotype classification frameworks. We also demonstrate that temporal information further improves the performance of the phenotype classification system.

plant biology↗

bioRxiv2017

Intersubject brain network organization during dynamic anxious anticipation

How do large-scale brain networks reorganize during the waxing and waning of anxious anticipation? Here, threat was dynamically modulated during functional MRI as two circles slowly meandered on the screen; if they touched, an unpleasant shock was delivered. We employed intersubject network analysis, which allows the investigation of network-level properties \"across brains,\" and sought to determine how network properties changed during periods of approach (circles moving closer) and periods of retreat (circles moving apart). Dynamic threat altered network cohesion across the salience, executive, and task-negative networks, as well as subcortical regions. Functional connections between subcortical regions and the salience network also increased during approach vs. retreat, including the putative periaqueductal gray, habenula, and amygdala, showing that the latter is involved under conditions of relatively prolonged and uncertain threat (the bed nucleus of the stria terminalis was observed during both approach and retreat). Together, our findings unraveled dynamic properties of large-scale networks across participants while threat levels varied continuously, and demonstrate the potential of characterizing emotional processing at the level of distributed networks.\n\nSignificance StatementUnderstanding the brain basis of anxious anticipation is important not only from a basic research perspective, but because aberrant responding to uncertain future negative events is believed to be central to anxiety disorders. Although previous studies have investigated how brain responses are sensitive to threat proximity, little is known about how patterns of response co-activation change during dynamic manipulations of threat. To address these important gaps in the literature, we studied the dynamics of emotional processing at the level of large-scale brain networks by devising a manipulation in which threat was dynamically modulated during functional MRI scanning.

neuroscience↗