Machine and deep learning models for accurate detection of ischemia and scar with myocardial blood flow positron emission tomography imaging.

BACKGROUND: Quantification of myocardial blood flow (MBF) is used for the noninvasive diagnosis of patients with coronary artery disease (CAD). This study compared traditional statistics, machine learning, and deep learning techniques in their ability to diagnose disease using only the rest and stress MBF values. METHODS: This study included 3245 rest and stress rubidium-82 positron emission tomography (PET) studies and matching diagnostic labels from perfusion reports. Standard logistic regression, lasso logistic regression, support vector machine, random forest, multilayer perceptron, and dense U-Net were compared for per-patient detection and per-vessel localization of scars and ischemia. RESULTS: Receiver-operator characteristic area under the curve (AUC) of machine learning models was significantly higher than those of traditional statistics models for per-patient detection of disease (0.92-0.95 vs. 0.87) but not for per-vessel localization of ischemia or scar. Random forest showed the highest AUC = 0.95 among the different models compared. On the final hold-out set for generalizability, random forest showed an AUC of 0.92 for detection and 0.89 for localization of perfusion abnormalities. CONCLUSIONS: For per-vessel localization, simple models trained on segmental data performed similarly to a convolutional neural network trained on polar-map data, highlighting the need to justify the use of complex predictive algorithms through comparison with simpler methods.

Mammalian Target of Rapamycin (mTOR) Tagging Promotes Dendritic Branch Variability through the Capture of Ca2+/Calmodulin-dependent Protein Kinase II α (CaMKIIα) mRNAs by the RNA-binding Protein HuD.

The fate of a memory, whether stored or forgotten, is determined by the ability of an active or tagged synapse to undergo changes in synaptic efficacy requiring protein synthesis of plasticity-related proteins. A synapse can be tagged, but without the "capture" of plasticity-related proteins, it will not undergo long lasting forms of plasticity (synaptic tagging and capture hypothesis). What the "tag" is and how plasticity-related proteins are captured at tagged synapses are unknown. Ca(2+)/calmodulin-dependent protein kinase II α (CaMKIIα) is critical in learning and memory and is synthesized locally in neuronal dendrites. The mechanistic (mammalian) target of rapamycin (mTOR) is a protein kinase that increases CaMKIIα protein expression; however, the mechanism and site of dendritic expression are unknown. Herein, we show that mTOR activity mediates the branch-specific expression of CaMKIIα, favoring one secondary, daughter branch over the other in a single neuron. mTOR inhibition decreased the dendritic levels of CaMKIIα protein and mRNA by shortening its poly(A) tail. Overexpression of the RNA-stabilizing protein HuD increased CaMKIIα protein levels and preserved its selective expression in one daughter branch over the other when mTOR was inhibited. Unexpectedly, deleting the third RNA recognition motif of HuD, the domain that binds the poly(A) tail, eliminated the branch-specific expression of CaMKIIα when mTOR was active. These results provide a model for one molecular mechanism that may underlie the synaptic tagging and capture hypothesis where mTOR is the tag, preventing deadenylation of CaMKIIα mRNA, whereas HuD captures and promotes its expression in a branch-specific manner.

Pushing the threshold: How NMDAR antagonists induce homeostasis through protein synthesis to remedy depression.

Healthy neurons have an optimal operating range, coded globally by the frequency of action potentials or locally by calcium. The maintenance of this range is governed by homeostatic plasticity. Here, we discuss how new approaches to treat depression alter synaptic activity. These approaches induce the neuron to recruit homeostatic mechanisms to relieve depression. Homeostasis generally implies that the direction of activity necessary to restore the neuron's critical operating range is opposite in direction to its current activity pattern. Unconventional antidepressant therapies-deep brain stimulation and NMDAR antagonists-alter the neuron's "depressed" state by pushing the neuron's current activity in the same direction but to the extreme edge. These therapies rally the intrinsic drive of neurons in the opposite direction, thereby allowing the cell to return to baseline activity, form new synapses, and restore proper communication. In this review, we discuss seminal studies on protein synthesis dependent homeostatic plasticity and their contribution to our understanding of molecular mechanisms underlying the effectiveness of NMDAR antagonists as rapid antidepressants. Rapid antidepressant efficacy is likely to require a cascade of mRNA translational regulation. Emerging evidence suggests that changes in synaptic strength or intrinsic excitability converge on the same protein synthesis pathways, relieving depressive symptoms. Thus, we address the question: Are there multiple homeostatic mechanisms that induce the neuron and neuronal circuits to self-correct to regulate mood in vivo? Targeting alternative ways to induce homeostatic protein synthesis may provide, faster, safer, and longer lasting antidepressants. This article is part of a Special Issue entitled SI:RNA Metabolism in Disease.

FMRP regulates an ethanol-dependent shift in GABABR function and expression with rapid antidepressant properties.

Alcohol promotes lasting neuroadaptive changes that may provide relief from depressive symptoms, often referred to as the self-medication hypothesis. However, the molecular/synaptic pathways that are shared by alcohol and antidepressants are unknown. In the current study, acute exposure to ethanol produced lasting antidepressant and anxiolytic behaviours. To understand the functional basis of these behaviours, we examined a molecular pathway that is activated by rapid antidepressants. Ethanol, like rapid antidepressants, alters γ-aminobutyric acid type B receptor (GABABR) expression and signalling, to increase dendritic calcium. Furthermore, new GABABRs are synthesized in response to ethanol treatment, requiring fragile-X mental retardation protein (FMRP). Ethanol-dependent changes in GABABR expression, dendritic signalling, and antidepressant efficacy are absent in Fmr1-knockout (KO) mice. These findings indicate that FMRP is an important regulator of protein synthesis following alcohol exposure, providing a molecular basis for the antidepressant efficacy of acute ethanol exposure.

Synaptic adaptations by alcohol and drugs of abuse: changes in microRNA expression and mRNA regulation.

Local translation of mRNAs is a mechanism by which cells can rapidly remodel synaptic structure and function. There is ample evidence for a role of synaptic translation in the neuroadaptations resulting from chronic drug use and abuse. Persistent and coordinated changes of many mRNAs, globally and locally, may have a causal role in complex disorders such as addiction. In this review we examine the evidence that translational regulation by microRNAs drives synaptic remodeling and mRNA expression, which may regulate the transition from recreational to compulsive drug use. microRNAs are small, non-coding RNAs that control the translation of mRNAs in the cell and within spatially restricted sites such as the synapse. microRNAs typically repress the translation of mRNAs into protein by binding to the 3'UTR of their targets. As 'master regulators' of many mRNAs, changes in microRNAs could account for the systemic alterations in mRNA and protein expression observed with drug abuse and dependence. Recent studies indicate that manipulation of microRNAs affects addiction-related behaviors such as the rewarding properties of cocaine, cocaine-seeking behavior, and self-administration rates of alcohol. There is limited evidence, however, regarding how synaptic microRNAs control local mRNA translation during chronic drug exposure and how this contributes to the development of dependence. Here, we discuss research supporting microRNA regulation of local mRNA translation and how drugs of abuse may target this process. The ability of synaptic microRNAs to rapidly regulate mRNAs provides a discrete, localized system that could potentially be used as diagnostic and treatment tools for alcohol and other addiction disorders.