Brain states in freely behaving marmosets

Study Objectives

We evaluated common marmosets as a perspective animal model to study human sleep and wake states.

Methods

Using wireless neurologger recordings, we performed longitudinal multichannel local field potential (LFP) cortical, hippocampal, neck muscle, and video recordings in three freely behaving marmosets. The brain states were formally identified using self-organizing maps.

Results

Marmosets were generally awake during the day with occasional 1–2 naps, and they slept during the night. Major electrographic patterns fall in five clearly distinguished categories: wakefulness, drowsiness, light and deep NREM sleep, and REM. Marmosets typically had 14–16 sleep cycles per night, with either gradually increasing or relatively low, but stable delta power within the cycle. Overall, the delta power decreased throughout the night sleep. Marmosets demonstrated prominent high amplitude somatosensory mu-rhythm (10–15 Hz), accompanied with neocortical ripples, and alternated with occipital alpha rhythm (10–15 Hz). NREM sleep was characterized with the presence of high amplitude slow waves, sleep spindles and ripples in neocortex, and sharp-wave-ripple complexes in CA1. Light and deep stages differed in levels of delta and sigma power and muscle tone. REM sleep was defined with low muscle tone and activated LFP with predominant beta-activity and rare spindle-like or mu-like events.

Conclusions

Multiple features of sleep–wake state distribution and electrographic patterns associated with behavioral states in marmosets closely match human states, although marmoset have shorter sleep cycles. This demonstrates that marmosets represent an excellent model to study origin of human electrographical rhythms and brain states.

Major areas of interest of artificial intelligence research applied to health care administrative data: a scoping review

Introduction: The ongoing collection of large medical data has created conditions for application of artificial intelligence (AI) in research. This scoping review aimed to identify major areas of interest of AI applied to health care administrative data.

Methods: The search was performed in seven databases: Medline, Embase, CINAHL, Web of science, IEEE, ICM digital library, and Compendex. We included articles published between January 2001 and March 2021, that described research with AI applied to medical diagnostics, pharmacotherapy, and health outcomes data. We screened the full text content and used natural language processing to automatically extract health areas of interest, principal AI methods, and names of medications.

Results: Out of 14,864 articles, 343 were included. We determined ten areas of interest, the most common being health diagnostic or treatment outcome prediction (32%); representation of medical data, clinical pathways, and data temporality (i.e., transformation of raw medical data into compact and analysis-friendly format) (22%); and adverse drug effects, drug-drug interactions, and medication cascades (15%). Less attention has been devoted to areas such as health effects of polypharmacy (1%); and reinforcement learning (1%). The most common AI methods were decision trees, cluster analysis, random forests, and support vector machines. Most frequently mentioned medications included insulin, metformin, vitamins, acetaminophen, and heparin.

Conclusions: The scoping review revealed the potential of AI application to health-related studies. However, several areas of interest in pharmacoepidemiology are sparsely reported, and the lack of details in studies related to pharmacotherapy suggests that AI could be used more optimally in pharmacoepidemiologic research.

Neuronal activities during palatable food consumption in the reward system of binge-like eating female rats

Binge eating disorder (BED), characterized by bingeing episodes and compulsivity, is the most prevalent eating disorder; however, little is known about its neurobiological underpinnings. In humans, BED is associated with desensitization of the reward system, specifically, the medial prefrontal cortex (mPFC), nucleus accumbens (Acb), and ventral tegmental area (VTA). Additionally, BED patients feel relieved during bingeing, suggesting that bingeing helps to decrease the negative emotions they were feeling prior to the binge episode. However, the mechanisms that underlie this feeling of relief in BED patients have not been well investigated. To investigate neuronal activity before and during palatable food consumption in BED, we performed in vivo electrophysiological recordings in a binge-like eating rat model (bingeing, n = 12 and non-bingeing, n = 14) and analyzed the firing rate of neurons in the mPFC, Acb, and VTA before and during access to sucrose solution. We also investigated changes in the firing rate of neurons in these regions during and between active bingeing, which may underlie the feeling of relief in BED patients. We found that neuronal firing rates of mPFC and VTA neurons in bingeing rats were lower than those in non-bingeing rats before and during sucrose consumption. Palatable food consumption increased neuronal firing rates during and between active bingeing in bingeing rats. Our results suggest a desynchronization in the activity of reward system regions, specifically in the mPFC, in bingeing rats, which may also contribute to BED. These results are consistent with those of functional magnetic resonance imaging (fMRI) studies that reported decreased activity in the reward system in BED patients. We propose that increased neuronal activity in the mPFC, Acb, or VTA produces an antidepressant effect in rats, which may underlie the sense of relief patients express during bingeing episodes.

Machine learning support for decision making in kidney transplantation: step-by-step development of a technological solution (Preprint)

Background

Kidney transplantation is the preferred treatment option for patients with end-stage renal disease. To maximize patient and graft survival, the allocation of donor organs to potential recipients requires careful consideration.

Objective

This study aimed to develop an innovative technological solution to enable better prediction of kidney transplant survival for each potential donor-recipient pair.

Methods

We used deidentified data on past organ donors, recipients, and transplant outcomes in the United States from the Scientific Registry of Transplant Recipients. To predict transplant outcomes for potential donor-recipient pairs, we used several survival analysis models, including regression analysis (Cox proportional hazards), random survival forests, and several artificial neural networks (DeepSurv, DeepHit, and recurrent neural network [RNN]). We evaluated the performance of each model in terms of its ability to predict the probability of graft survival after kidney transplantation from deceased donors. Three metrics were used: the C-index, integrated Brier score, and integrated calibration index, along with calibration plots.

Results

On the basis of the C-index metrics, the neural network–based models (DeepSurv, DeepHit, and RNN) had better discriminative ability than the Cox model and random survival forest model (0.650, 0.661, and 0.659 vs 0.646 and 0.644, respectively). The proposed RNN model offered a compromise between the good discriminative ability and calibration and was implemented in a technological solution of technology readiness level 4.

Conclusions

Our technological solution based on the RNN model can effectively predict kidney transplant survival and provide support for medical professionals and candidate recipients in determining the most optimal donor-recipient pair.

Exploring polypharmacy with artificial intelligence: data analysis protocol

BACKGROUND

Polypharmacy is common among older adults and it represents a public health concern, due to the negative health impacts potentially associated with the use of several medications. However, the large number of medication combinations and sequences of use makes it complicated for traditional statistical methods to predict which therapy is genuinely associated with health outcomes. The project aims to use artificial intelligence (AI) to determine the quality of polypharmacy among older adults with chronic diseases in the province of Québec, Canada.

METHODS

We will use data from the Quebec Integrated Chronic Disease Surveillance System (QICDSS). QICDSS contains information about prescribed medications in older adults in Quebec collected over 20 years. It also includes diagnostic codes and procedures, and sociodemographic data linked through a unique identification number for each individual. Our research will be structured around three interconnected research axes: AI, Health, and Law&Ethics. The AI research axis will develop algorithms for finding frequent patterns of medication use that correlate with health events, considering data locality and temporality (explainable AI or XAI). The Health research axis will translate these patterns into polypharmacy indicators relevant to public health surveillance and clinicians. The Law&Ethics axis will assess the social acceptability of the algorithms developed using AI tools and the indicators developed by the Heath axis and will ensure that the developed indicators neither discriminate against any population group nor increase the disparities already present in the use of medications.

DISCUSSION

The multi-disciplinary research team consists of specialists in AI, health data, statistics, pharmacy, public health, law, and ethics, which will allow investigation of polypharmacy from different points of view and will contribute to a deeper understanding of the clinical, social, and ethical issues surrounding polypharmacy and its surveillance, as well as the use of AI for health record data. The project results will be disseminated to the scientific community, healthcare professionals, and public health decision-makers in peer-reviewed publications, scientific meetings, and reports. The diffusion of the results will ensure the confidentiality of individual data.

0354 Age-Related Spectral Changes in NREM And REM Sleep in Mice are Global and Not Local

Introduction

Brain topography modulates age-related changes in the human sleep electroencephalogram, which are linked with differences in integrity of specific cortical areas and may reflect local changes in sleep homeostasis. In mice, there is conflicting evidence regarding the topography of age-related changes for NREM and REM sleep. To disambiguate this issue, we investigated in mice the topography of age-related spectral differences for REM and NREM sleep.

Methods

LFP electrodes were implanted in 5 cortical areas and in the hippocampus of 17 C57/BL6 mice (8 young and 9 old, mean age = 7.5 and 16 months). Mice LFPs were recorded for a week and states of vigilance were semi-automatically detected in light and dark periods (12h-12h). Spectral analysis was run on 4s windows. Values were averaged for each electrode and in each period of the light/dark cycle in REM/NREM sleep for slow delta (0.25-2Hz), delta (2-4Hz), theta (4-8Hz), sigma (10-16Hz) and ripples (150-200Hz). Mixed models were computed separately for REM and NREM in dark and light period, with age as group factor and electrode and frequency as repeated factors.

Results

Two-way interactions were found between age and frequency and between electrode and frequency, for NREM and REM in dark and light periods. Each frequency band, except ripples, showed a topographical signature in NREM and REM (e.g. higher power in anterior compared to posterior areas for delta band in NREM sleep). These relative patterns did not change in older mice, but global changes occurred on all electrodes: in older mice, delta power was globally higher in NREM and REM sleep whereas sigma power was lower in REM sleep.

Conclusion

Age-related changes in spectral power of sleeping mice do not vary according to brain topography as in humans. Sleep deprivation studies are needed to investigate whether age is associated with global changes in sleep homeostasis in mice.

Support

This work has been supported by the Quebec Fonds de Recherche Nature et Technologies (FQRNT).

Sleep-Wake Cycle in Young and Older Mice

Sleep plays a key role in multiple cognitive functions and sleep pattern changes with aging. Human studies revealed that aging decreases sleep efficiency and reduces the total sleep time, the time spent in slow-wave sleep (SWS), and the delta power (1–4 Hz) during sleep; however, some studies of sleep and aging in mice reported opposing results. The aim of our work is to estimate how features of sleep–wake state in mice during aging could correspond to age-dependent changes observed in human. In this study, we investigated the sleep/wake cycle in young (3 months old) and older (12 months old) C57BL/6 mice using local-field potentials (LFPs). We found that older adult mice sleep more than young ones but only during the dark phase of sleep-wake cycle. Sleep fragmentation and sleep during the active phase (dark phase of cycle), homologous to naps, were higher in older mice. Older mice show a higher delta power in frontal cortex, which was accompanied with similar trend for age differences in slow wave density. We also investigated regional specificity of sleep–wake electrographic activities and found that globally posterior regions of the cortex show more rapid eye movement (REM) sleep whereas somatosensory cortex displays more often SWS patterns. Our results indicate that the effects of aging on the sleep–wake activities in mice occur mainly during the dark phase and the electrode location strongly influence the state detection. Despite some differences in sleep–wake cycle during aging between human and mice, some features of mice sleep share similarity with human sleep during aging.

Slow wave detection in sleeping mice: Comparison of traditional and machine learning methods

BACKGROUND

During slow-wave sleep the electroencephalographic (EEG) and local field potential (LFP) recordings reveal the presence of large amplitude slow waves. Systematic extraction of individual slow waves is not trivial.

NEW METHOD

In this study, we used the neural network pattern recognition to detect individual slow waves in LFP recorded from mice as well as other commonly used methods that are based on fast frequencies modulation, amplitude, or duration.

RESULTS

The number and quality of events detected as slow waves depended on the chosen method of detection, level of thresholds, or on combination of methods. Each individual method yields some false-positive and false-negative detections. Typically, the fast frequency-method has a higher false discovery rate, but almost no missing waves; amplitude-based method has relatively high false-positive and false-negative rates; duration-based method has low false-negative rates; neural network pattern recognition approach has the lowest false-positive rate among individual methods, often rejecting waves that were falsely detected by other approaches. Combining all 4 detection methods practically eliminated false-positive errors, but a large number of slow waves remained undetected.

CONCLUSIONS

The use of a particular method of slow wave detection needs to be adjusted to the objectives of a given study: to detect all slow waves, but also numerous false positives can be achieved using the fast frequency approach. Neural network pattern recognition method alone can detect slow waves with the lowest false-positive rate, that can be further minimized with the use of combination of other methods.