Dropout in Neural Networks Simulates the Paradoxical Effects of Deep Brain Stimulation on Memory

OGNNMDA: a computational model for microbe-drug association prediction based on ordered message-passing graph neural networks

In recent years, many excellent computational models have emerged in microbe-drug association prediction, but their performance still has room for improvement. This paper proposed the OGNNMDA framework, which applied an ordered message-passing mechanism to distinguish the different neighbor information in each message propagation layer, and it achieved a better embedding ability through deeper network layers. Firstly, the method calculates four similarity matrices based on microbe functional similarity, drug chemical structure similarity, and their respective Gaussian interaction profile kernel similarity. After integrating these similarity matrices, it concatenates the integrated similarity matrix with the known association matrix to obtain the microbe-drug heterogeneous matrix. Secondly, it uses a multi-layer ordered message-passing graph neural network encoder to encode the heterogeneous network and the known association information adjacency matrix, thereby obtaining the final embedding features of the microbe-drugs. Finally, it inputs the embedding features into the bilinear decoder to get the final prediction results. The OGNNMDA method performed comparative experiments, ablation experiments, and case studies on the aBiofilm, MDAD and DrugVirus datasets using 5-fold cross-validation. The experimental results showed that OGNNMDA showed the strongest prediction performance on aBiofilm and MDAD and obtained sub-optimal results on DrugVirus. In addition, the case studies on well-known drugs and microbes also support the effectiveness of the OGNNMDA method. Source codes and data are available at: https://github.com/yyzg/OGNNMDA.

Distinctive properties of biological neural networks and recent advances in bottom-up approaches toward a better biologically plausible neural network

Although it may appear infeasible and impractical, building artificial intelligence (AI) using a bottom-up approach based on the understanding of neuroscience is straightforward. The lack of a generalized governing principle for biological neural networks (BNNs) forces us to address this problem by converting piecemeal information on the diverse features of neurons, synapses, and neural circuits into AI. In this review, we described recent attempts to build a biologically plausible neural network by following neuroscientifically similar strategies of neural network optimization or by implanting the outcome of the optimization, such as the properties of single computational units and the characteristics of the network architecture. In addition, we proposed a formalism of the relationship between the set of objectives that neural networks attempt to achieve, and neural network classes categorized by how closely their architectural features resemble those of BNN. This formalism is expected to define the potential roles of top-down and bottom-up approaches for building a biologically plausible neural network and offer a map helping the navigation of the gap between neuroscience and AI engineering.

Prelimbic Cortical Stimulation with L-methionine Enhances Cognition through Hippocampal DNA Methylation and Neuroplasticity Mechanisms

Declining global DNA methylation and cognitive impairment are reported to occur in the normal aging process. It is not known if DNA methylation plays a role in the efficacy of memory-enhancing therapies. In this study, aged animals were administered prelimbic cortical deep brain stimulation (PrL DBS) and/or L-methionine (MET) treatment. We found that PrL DBS and MET (MET-PrL DBS) co-administration resulted in hippocampal-dependent spatial memory enhancements in aged animals. Molecular data suggested MET-PrL DBS induced DNA methyltransferase DNMT3a-dependent methylation, robust synergistic upregulation of neuroplasticity-related genes, and simultaneous inhibition of the memory-suppressing gene calcineurin in the hippocampus. We further found that MET-PrL DBS also activated the PKA-CaMKIIα-BDNF pathway, increased hippocampal neurogenesis, and enhanced dopaminergic and serotonergic neurotransmission. We next inhibited the activity of DNA methyltransferase (DNMT) by RG108 infusion in the hippocampus of young animals to establish a causal relationship between DNMT activity and the effects of PrL DBS. Hippocampal DNMT inhibition in young animals was sufficient to recapitulate the behavioral deficits observed in aged animals and abolished the memory-enhancing and molecular effects of PrL DBS. Our findings implicate hippocampal DNMT as a therapeutic target for PrL DBS and pave way for the potential use of non-invasive neuromodulation modalities against dementia.

Entropic associative memory for manuscript symbols

Manuscript symbols can be stored, recognized and retrieved from an entropic digital memory that is associative and distributed but yet declarative; memory retrieval is a constructive operation, memory cues to objects not contained in the memory are rejected directly without search, and memory operations can be performed through parallel computations. Manuscript symbols, both letters and numerals, are represented in Associative Memory Registers that have an associated entropy. The memory recognition operation obeys an entropy trade-off between precision and recall, and the entropy level impacts on the quality of the objects recovered through the memory retrieval operation. The present proposal is contrasted in several dimensions with neural networks models of associative memory. We discuss the operational characteristics of the entropic associative memory for retrieving objects with both complete and incomplete information, such as severe occlusions. The experiments reported in this paper add evidence on the potential of this framework for developing practical applications and computational models of natural memory.

Transcorneal electrical stimulation enhances cognitive functions in aged and 5XFAD mouse models

Dementia is a major burden on global health for which there are no effective treatments. The use of noninvasive visual stimulation to ameliorate cognitive deficits is a novel concept that may be applicable for treating dementia. In this study, we investigated the effects of transcorneal electrical stimulation (TES) on memory enhancement using two mouse models, in aged mice and in the 5XFAD model of Alzheimer's disease. After 3 weeks of TES treatment, mice were subjected to Y-maze and Morris water maze tests to assess hippocampal-dependent learning and memory. Immunostaining of the hippocampus of 5XFAD mice was also performed to examine the effects of TES on amyloid plaque pathology. The results showed that TES improved the performance of both aged and 5XFAD mice in memory tests. TES also reduced hippocampal plaque deposition in male, but not female, 5XFAD mice. Moreover, TES significantly reversed the downregulated level of postsynaptic protein 95 in the hippocampus of male 5XFAD mice, suggesting the effects of TES involve a postsynaptic mechanism. Overall, these findings support further investigation of TES as a potential treatment for cognitive dysfunction and mechanistic studies of TES effects in other dementia models.

Therapeutic Potential of Human Stem Cell Implantation in Alzheimer's Disease

Alzheimer's disease (AD) is a progressive debilitating neurodegenerative disease and the most common form of dementia in the older population. At present, there is no definitive effective treatment for AD. Therefore, researchers are now looking at stem cell therapy as a possible treatment for AD, but whether stem cells are safe and effective in humans is still not clear. In this narrative review, we discuss both preclinical studies and clinical trials on the therapeutic potential of human stem cells in AD. Preclinical studies have successfully differentiated stem cells into neurons in vitro, indicating the potential viability of stem cell therapy in neurodegenerative diseases. Preclinical studies have also shown that stem cell therapy is safe and effective in improving cognitive performance in animal models, as demonstrated in the Morris water maze test and novel object recognition test. Although few clinical trials have been completed and many trials are still in phase I and II, the initial results confirm the outcomes of the preclinical studies. However, limitations like rejection, tumorigenicity, and ethical issues are still barriers to the advancement of stem cell therapy. In conclusion, the use of stem cells in the treatment of AD shows promise in terms of effectiveness and safety.

Prelimbic cortical stimulation disrupts fear memory consolidation through ventral hippocampal dopamine D2 receptors

BACKGROUND AND PURPOSE

Anxiety disorders pose one of the biggest threats to mental health worldwide, yet current therapeutics have been mostly ineffective due to issues with relapse, efficacy and toxicity of the medications. Deep brain stimulation (DBS) is a promising therapy for treatment-resistant psychiatric disorders including anxiety, but very little is known about the effects of deep brain stimulation on fear memories.

EXPERIMENTAL APPROACH

In this study, we employed a standard tone-footshock fear conditioning paradigm and modified plus maze discriminative avoidance task to probe the effects of prelimbic cortex deep brain stimulation on various stages of memory.

KEY RESULTS

We identified memory consolidation stage as a critical time point to disrupt fear memory via prelimbic cortex deep brain stimulation. The observed disruption was partially modulated by the inactivation of the ventral hippocampus and the transient changes in ventral hippocampus dopamine (D₂ ) receptors expression upon prelimbic cortex deep brain stimulation. We also observed wide-scale changes of various neurotransmitters and their metabolites in ventral hippocampus, confirming its important role in response to prelimbic cortex deep brain stimulation.

CONCLUSION AND IMPLICATIONS

These findings highlight the molecular mechanism in the ventral hippocampus in response to prelimbic cortex stimulation and may have translational value, indicating that targeting the prelimbic cortex in the memory consolidation stage via non-invasive neuromodulation techniques may be a feasible therapeutic strategy against anxiety disorders.

Prelimbic Cortical Stimulation Improves Spatial Memory Through Distinct Patterns of Hippocampal Gene Expression in Aged Rats

Dementia poses major health challenges worldwide, yet current treatments are faced with issues of efficacy and toxicity. Deep brain stimulation (DBS) is a promising non-pharmacological treatment for dementia, but most DBS studies use young healthy animals, which may not be aetiologically relevant. In this study, we used an aged rat model in which cognitive decline occurs through a natural ageing process. We used a Morris water maze (MWM) to determine the effects of prelimbic cortex (PrL) DBS on memory in aged rats. To investigate the underlying mechanisms of the effects of DBS, we carried out microarray, quantitative PCR analysis, and mass spectrometry to detect gene expression and neurotransmitter changes in the hippocampus. We showed PrL DBS improved the performance in MWM, with related distinct patterns of gene expression involving G protein-coupled receptor pathways. We further found neurotransmitter changes in the dorsal hippocampus, which corroborated and extended the microarray findings. Our results suggest that non-neurogenesis pathways play roles in the effects of DBS. Further studies are needed to investigate the effects of DBS on memory beyond neurogenesis and to consider the highlighted pathways suggested by our data.