A pilot study examining the impact of lithium treatment and responsiveness on mnemonic discrimination in bipolar disorder

INTRODUCTION

Mnemonic discrimination (MD), the ability to discriminate new stimuli from similar memories, putatively involves dentate gyrus pattern separation. Since lithium may normalize dentate gyrus functioning in lithium-responsive bipolar disorder (BD), we hypothesized that lithium treatment would be associated with better MD in lithium-responsive BD patients.

METHODS

BD patients (N = 69; NResponders = 16 [23 %]) performed the Continuous Visual Memory Test (CVMT), which requires discriminating between novel and previously seen images. Before testing, all patients had prophylactic lithium responsiveness assessed over ≥1 year of therapy (with the Alda Score), although only thirty-eight patients were actively prescribed lithium at time of testing (55 %; 12/16 responders, 26/53 nonresponders). We then used computational modelling to extract patient-specific MD indices. Linear models were used to test how (A) lithium treatment, (B) lithium responsiveness via the continuous Alda score, and (C) their interaction, affected MD.

RESULTS

Superior MD performance was associated with lithium treatment exclusively in lithium-responsive patients (Lithium x AldaScore β = 0.257 [SE 0.078], p = 0.002). Consistent with prior literature, increased age was associated with worse MD (β = -0.03 [SE 0.01], p = 0.005).

LIMITATIONS

Secondary pilot analysis of retrospectively collected data in a cross-sectional design limits generalizability.

CONCLUSION

Our study is the first to examine MD performance in BD. Lithium is associated with better MD performance only in lithium responders, potentially due to lithium's effects on dentate gyrus granule cell excitability. Our results may influence the development of behavioural probes for dentate gyrus neuronal hyperexcitability in BD.

Granule cells perform frequency-dependent pattern separation in a computational model of the dentate gyrus

Mnemonic discrimination (MD) may be dependent on oscillatory perforant path input frequencies to the hippocampus in a "U"-shaped fashion, where some studies show that slow and fast input frequencies support MD, while other studies show that intermediate frequencies disrupt MD. We hypothesize that pattern separation (PS) underlies frequency-dependent MD performance. We aim to study, in a computational model of the hippocampal dentate gyrus (DG), the network and cellular mechanisms governing this putative "U"-shaped PS relationship. We implemented a biophysical model of the DG that produces the hypothesized "U"-shaped input frequency-PS relationship, and its associated oscillatory electrophysiological signatures. We subsequently evaluated the network's PS ability using an adapted spatiotemporal task. We undertook systematic lesion studies to identify the network-level mechanisms driving the "U"-shaped input frequency-PS relationship. A minimal circuit of a single granule cell (GC) stimulated with oscillatory inputs was also used to study potential cellular-level mechanisms. Lesioning synapses onto GCs did not impact the "U"-shaped input frequency-PS relationship. Furthermore, GC inhibition limits PS performance for fast frequency inputs, while enhancing PS for slow frequency inputs. GC interspike interval was found to be input frequency dependent in a "U"-shaped fashion, paralleling frequency-dependent PS observed at the network level. Additionally, GCs showed an attenuated firing response for fast frequency inputs. We conclude that independent of network-level inhibition, GCs may intrinsically be capable of producing a "U"-shaped input frequency-PS relationship. GCs may preferentially decorrelate slow and fast inputs via spike timing reorganization and high frequency filtering.

Mechanisms of Sustained Increases in γ Power Post-Ketamine in a Computational Model of the Hippocampal CA3: Implications for Ketamine's Antidepressant Mechanism of Action

Subanaesthetic doses of ketamine increase γ oscillation power in neural activity measured using electroencephalography (EEG), and this effect lasts several hours after ketamine administration. The mechanisms underlying this effect are unknown. Using a computational model of the hippocampal cornu ammonis 3 (CA3) network, which is known to reproduce ketamine's acute effects on γ power, we simulated the plasticity of glutamatergic synapses in pyramidal cells to test which of the following hypotheses would best explain this sustained γ power: the direct inhibition hypothesis, which proposes that increased γ power post-ketamine administration may be caused by the potentiation of recurrent collateral synapses, and the disinhibition hypothesis, which proposes that potentiation affects synapses from both recurrent and external inputs. Our results suggest that the strengthening of external connections to pyramidal cells is able to account for the sustained γ power increase observed post-ketamine by increasing the overall activity of and synchrony between pyramidal cells. The strengthening of recurrent pyramidal weights, however, would cause an additional phase shifted voltage increase that ultimately reduces γ power due to partial cancellation. Our results therefore favor the disinhibition hypothesis for explaining sustained γ oscillations after ketamine administration.

A critical evaluation of dynamical systems models of bipolar disorder

Bipolar disorder (BD) is a mood disorder involving recurring (hypo)manic and depressive episodes. The inherently temporal nature of BD has inspired its conceptualization using dynamical systems theory, which is a mathematical framework for understanding systems that evolve over time. In this paper, we provide a critical review of the dynamical systems models of BD. Owing to the heterogeneity of methodological and experimental designs in computational modeling, we designed a structured approach that parallels the appraisal of animal models by their face, predictive, and construct validity. This tool, the validity appraisal guide for computational models (VAG-CM), is not an absolute measure of validity, but rather a guide for a more objective appraisal of models in this review. We identified 26 studies published before November 18, 2021 that proposed generative dynamical systems models of time-varying signals in BD. Two raters independently applied the VAG-CM to the included studies, obtaining a mean Cohen's κ of 0.55 (95% CI [0.45, 0.64]) prior to establishing consensus ratings. Consensus VAG-CM ratings revealed three model/study clusters: data-driven models with face validity, theory-driven models with predictive validity, and theory-driven models lacking all forms of validity. We conclude that future modeling studies should employ a hybrid approach that first operationalizes BD features of interest using empirical data to achieve face validity, followed by explanations of those features using generative models with components that are homologous to physiological or psychological systems involved in BD, to achieve construct validity. Such models would be best developed alongside long-term prospective cohort studies involving a collection of multimodal time-series data. We also encourage future studies to extend, modify, and evaluate the VAG-CM approach for a wider breadth of computational modeling studies and psychiatric disorders.

Machine learning for contour classification in TG-263 noncompliant databases

A large volume of medical data are labeled using nonstandardized nomenclature. Although efforts have been made by the American Association of Physicists in Medicine (AAPM) to standardize nomenclature through Task Group 263 (TG-263), there remain noncompliant databases. This work aims to create an algorithm that can analyze anatomical contours in patients with head and neck cancer and classify them into TG-263 compliant nomenclature. To create an accurate algorithm capable of such classification, a combined approaching using both binary images of individual slices of anatomical contours themselves, as well as center of mass coordinates of the structures are input into a neural network. The center of mass coordinates were scaled using two normalization schemes, a simple linear normalization scheme agnostic of the patient anatomy, and an anatomical normalization scheme dependent on patient anatomy. The results of all of the individual slice classifications are then aggregated into a single classification by means of a voting algorithm. The total classification accuracy of the final algorithms was 97.6% mean accuracy per class for nonanatomically normalization scheme, and 97.9% mean accuracy per class for anatomically normalization scheme. The total accuracy was 99.0% (13 errors in 1302 structures) for the nonanatomically normalization scheme, and 98.3% (22 errors in 1302 structures) for the anatomically normalization scheme.

Plankton classification with high-throughput submersible holographic microscopy and transfer learning

BACKGROUND

Plankton are foundational to marine food webs and an important feature for characterizing ocean health. Recent developments in quantitative imaging devices provide in-flow high-throughput sampling from bulk volumes-opening new ecological challenges exploring microbial eukaryotic variation and diversity, alongside technical hurdles to automate classification from large datasets. However, a limited number of deployable imaging instruments have been coupled with the most prominent classification algorithms-effectively limiting the extraction of curated observations from field deployments. Holography offers relatively simple coherent microscopy designs with non-intrusive 3-D image information, and rapid frame rates that support data-driven plankton imaging tasks. Classification benchmarks across different domains have been set with transfer learning approaches, focused on repurposing pre-trained, state-of-the-art deep learning models as classifiers to learn new image features without protracted model training times. Combining the data production of holography, digital image processing, and computer vision could improve in-situ monitoring of plankton communities and contribute to sampling the diversity of microbial eukaryotes.

RESULTS

Here we use a light and portable digital in-line holographic microscope (The HoloSea) with maximum optical resolution of 1.5 μm, intensity-based object detection through a volume, and four different pre-trained convolutional neural networks to classify > 3800 micro-mesoplankton (> 20 μm) images across 19 classes. The maximum classifier performance was quickly achieved for each convolutional neural network during training and reached F1-scores > 89%. Taking classification further, we show that off-the-shelf classifiers perform strongly across every decision threshold for ranking a majority of the plankton classes.

CONCLUSION

These results show compelling baselines for classifying holographic plankton images, both rare and plentiful, including several dinoflagellate and diatom groups. These results also support a broader potential for deployable holographic microscopes to sample diverse microbial eukaryotic communities, and its use for high-throughput plankton monitoring.

Prediction of lithium response using genomic data

Predicting lithium response prior to treatment could both expedite therapy and avoid exposure to side effects. Since lithium responsiveness may be heritable, its predictability based on genomic data is of interest. We thus evaluate the degree to which lithium response can be predicted with a machine learning (ML) approach using genomic data. Using the largest existing genomic dataset in the lithium response literature (n = 2210 across 14 international sites; 29% responders), we evaluated the degree to which lithium response could be predicted based on 47,465 genotyped single nucleotide polymorphisms using a supervised ML approach. Under appropriate cross-validation procedures, lithium response could be predicted to above-chance levels in two constituent sites (Halifax, Cohen’s kappa 0.15, 95% confidence interval, CI [0.07, 0.24]; and Würzburg, kappa 0.2 [0.1, 0.3]). Variants with shared importance in these models showed over-representation of postsynaptic membrane related genes. Lithium response was not predictable in the pooled dataset (kappa 0.02 [− 0.01, 0.04]), although non-trivial performance was achieved within a restricted dataset including only those patients followed prospectively (kappa 0.09 [0.04, 0.14]). Genomic classification of lithium response remains a promising but difficult task. Classification performance could potentially be improved by further harmonization of data collection procedures.

Measuring heterogeneity in normative models as the effective number of deviation patterns

Normative modeling is an increasingly popular method for characterizing the ways in which clinical cohorts deviate from a reference population, with respect to one or more biological features. In this paper, we extend the normative modeling framework with an approach for measuring the amount of heterogeneity in a cohort. This heterogeneity measure is based on the Representational Rényi Heterogeneity method, which generalizes diversity measurement paradigms used across multiple scientific disciplines. We propose that heterogeneity in the normative modeling setting can be measured as the effective number of deviation patterns; that is, the effective number of coherent patterns by which a sample of data differ from a distribution of normative variation. We show that lower effective number of deviation patterns is associated with the presence of systematic differences from a (non-degenerate) normative distribution. This finding is shown to be consistent across (A) application of a Gaussian process model to synthetic and real-world neuroimaging data, and (B) application of a variational autoencoder to well-understood database of handwritten images.

The definition and measurement of heterogeneity

Heterogeneity is an important concept in psychiatric research and science more broadly. It negatively impacts effect size estimates under case-control paradigms, and it exposes important flaws in our existing categorical nosology. Yet, our field has no precise definition of heterogeneity proper. We tend to quantify heterogeneity by measuring associated correlates such as entropy or variance: practices which are akin to accepting the radius of a sphere as a measure of its volume. Under a definition of heterogeneity as the degree to which a system deviates from perfect conformity, this paper argues that its proper measure roughly corresponds to the size of a system's event/sample space, and has units known as numbers equivalent. We arrive at this conclusion through focused review of more than 100 years of (re)discoveries of indices by ecologists, economists, statistical physicists, and others. In parallel, we review psychiatric approaches for quantifying heterogeneity, including but not limited to studies of symptom heterogeneity, microbiome biodiversity, cluster-counting, and time-series analyses. We argue that using numbers equivalent heterogeneity measures could improve the interpretability and synthesis of psychiatric research on heterogeneity. However, significant limitations must be overcome for these measures-largely developed for economic and ecological research-to be useful in modern translational psychiatric science.

Multiplicative Decomposition of Heterogeneity in Mixtures of Continuous Distributions

A system's heterogeneity (diversity) is the effective size of its event space, and can be quantified using the Rényi family of indices (also known as Hill numbers in ecology or Hannah-Kay indices in economics), which are indexed by an elasticity parameter q≥0. Under these indices, the heterogeneity of a composite system (the γ-heterogeneity) is decomposable into heterogeneity arising from variation within and between component subsystems (the α- and β-heterogeneity, respectively). Since the average heterogeneity of a component subsystem should not be greater than that of the pooled system, we require that γ≥α. There exists a multiplicative decomposition for Rényi heterogeneity of composite systems with discrete event spaces, but less attention has been paid to decomposition in the continuous setting. We therefore describe multiplicative decomposition of the Rényi heterogeneity for continuous mixture distributions under parametric and non-parametric pooling assumptions. Under non-parametric pooling, the γ-heterogeneity must often be estimated numerically, but the multiplicative decomposition holds such that γ≥α for q>0. Conversely, under parametric pooling, γ-heterogeneity can be computed efficiently in closed-form, but the γ≥α condition holds reliably only at q=1. Our findings will further contribute to heterogeneity measurement in continuous systems.

Prediction of lithium response using clinical data

OBJECTIVE

Promptly establishing maintenance therapy could reduce morbidity and mortality in patients with bipolar disorder. Using a machine learning approach, we sought to evaluate whether lithium responsiveness (LR) is predictable using clinical markers.

METHOD

Our data are the largest existing sample of direct interview-based clinical data from lithium-treated patients (n = 1266, 34.7% responders), collected across seven sites, internationally. We trained a random forest model to classify LR-as defined by the previously validated Alda scale-against 180 clinical predictors.

RESULTS

Under appropriate cross-validation procedures, LR was predictable in the pooled sample with an area under the receiver operating characteristic curve of 0.80 (95% CI 0.78-0.82) and a Cohen kappa of 0.46 (0.4-0.51). The model demonstrated a particularly low false-positive rate (specificity 0.91 [0.88-0.92]). Features related to clinical course and the absence of rapid cycling appeared consistently informative.

CONCLUSION

Clinical data can inform out-of-sample LR prediction to a potentially clinically relevant degree. Despite the relevance of clinical course and the absence of rapid cycling, there was substantial between-site heterogeneity with respect to feature importance. Future work must focus on improving classification of true positives, better characterizing between- and within-site heterogeneity, and further testing such models on new external datasets.

Asymmetrical reliability of the Alda score favours a dichotomous representation of lithium responsiveness

The Alda score is commonly used to quantify lithium responsiveness in bipolar disorder. Most often, this score is dichotomized into “responder” and “non-responder” categories, respectively. This practice is often criticized as inappropriate, since continuous variables are thought to invariably be “more informative” than their dichotomizations. We therefore investigated the degree of informativeness across raw and dichotomized versions of the Alda score, using data from a published study of the scale’s inter-rater reliability (n = 59 raters of 12 standardized vignettes each). After learning a generative model for the relationship between observed and ground truth scores (the latter defined by a consensus rating of the 12 vignettes), we show that the dichotomized scale is more robust to inter-rater disagreement than the raw 0-10 scale. Further theoretical analysis shows that when a measure’s reliability is stronger at one extreme of the continuum—a scenario which has received little-to-no statistical attention, but which likely occurs for the Alda score ≥ 7—dichotomization of a continuous variable may be more informative concerning its ground truth value, particularly in the presence of noise. Our study suggests that research employing the Alda score of lithium responsiveness should continue using the dichotomous definition, particularly when data are sampled across multiple raters.

Model-free prostate cancer segmentation from dynamic contrast-enhanced MRI with recurrent convolutional networks: A feasibility study

Dynamic contrast enhanced (DCE) magnetic resonance imaging (MRI) is a method of temporal imaging that is commonly used to aid in prostate cancer (PCa) diagnosis and staging. Typically, machine learning models designed for the segmentation and detection of PCa will use an engineered scalar image called K^trans to summarize the information in the DCE time-series images. This work proposes a new model that amalgamates the U-net and the convGRU neural network architectures for the purpose of interpreting DCE time-series in a temporal and spatial basis for segmenting PCa in MR images. Ultimately, experiments show that the proposed model using the DCE time-series images can outperform a baseline U-net segmentation model using K^trans. However, when other types of scalar MR images are considered by the models, no significant advantage is observed for the proposed model.

Modeling Saccadic Action Selection: Cortical and Basal Ganglia Signals Coalesce in the Superior Colliculus

The distributed nature of information processing in the brain creates a complex variety of decision making behavior. Likewise, computational models of saccadic decision making behavior are numerous and diverse. Here we present a generative model of saccadic action selection in the context of competitive decision making in the superior colliculus (SC) in order to investigate how independent neural signals may converge to interact and guide saccade selection, and to test if systematic variations can better replicate the variability in responses that are part of normal human behavior. The model was tasked with performing pro- and anti-saccades in order to replicate specific attributes of healthy human saccade behavior. Participants (ages 18-39) were instructed to either look toward (pro-saccade, well-practiced automated response) or away from (anti-saccade, combination of inhibitory and voluntary responses) a peripheral visual stimulus. They generated express and regular latency saccades in the pro-saccade task. In the anti-saccade task, correct reaction times were longer and participants occasionally looked at the stimulus (direction error) at either express or regular latencies. To gain a better understanding of the underlying neural processes that lead to saccadic action selection and response inhibition, we implemented 8 inputs inspired by systems neuroscience. These inputs reflected known sensory, automated, voluntary, and inhibitory components of cortical and basal ganglia activity that coalesces in the intermediate layers of the SC (SCi). The model produced bimodal reaction time distributions, where express and regular latency saccades had distinct modes, for both correct pro-saccades and direction errors in the anti-saccade task. Importantly, express and regular latency direction errors resulted from interactions of different inputs in the model. Express latency direction errors were due to a lack of pre-emptive fixation and inhibitory activity, which aloud sensory and automated inputs to initiate a stimulus-driven saccade. Regular latency errors occurred when the automated motor signals were stronger than the voluntary motor signals. While previous models have emulated fewer aspects of these behavioral findings, the focus of the simulations here is on the interaction of a wide variety of physiologically-based information integration producing a richer set of natural behavioral variability.

Learning-regulated context relevant topographical map

Kohonen's self-organizing map (SOM) is used to map high-dimensional data into a low-dimensional representation (typically a 2-D or 3-D space) while preserving their topological characteristics. A major reason for its application is to be able to visualize data while preserving their relation in the high-dimensional input data space as much as possible. Here, we are seeking to go further by incorporating semantic meaning in the low-dimensional representation. In a conventional SOM, the semantic context of the data, such as class labels, does not have any influence on the formation of the map. As an abstraction of neural function, the SOM models bottom-up self-organization but not feedback modulation which is also ubiquitous in the brain. In this paper, we demonstrate a hierarchical neural network, which learns a topographical map that also reflects the semantic context of the data. Our method combines unsupervised, bottom-up topographical map formation with top-down supervised learning. We discuss the mathematical properties of the proposed hierarchical neural network and demonstrate its abilities with empirical experiments.

A biological mechanism for Bayesian feature selection: Weight decay and raising the LASSO

Biological systems are capable of learning that certain stimuli are valuable while ignoring the many that are not, and thus perform feature selection. In machine learning, one effective feature selection approach is the least absolute shrinkage and selection operator (LASSO) form of regularization, which is equivalent to assuming a Laplacian prior distribution on the parameters. We review how such Bayesian priors can be implemented in gradient descent as a form of weight decay, which is a biologically plausible mechanism for Bayesian feature selection. In particular, we describe a new prior that offsets or "raises" the Laplacian prior distribution. We evaluate this alongside the Gaussian and Cauchy priors in gradient descent using a generic regression task where there are few relevant and many irrelevant features. We find that raising the Laplacian leads to less prediction error because it is a better model of the underlying distribution. We also consider two biologically relevant online learning tasks, one synthetic and one modeled after the perceptual expertise task of Krigolson et al. (2009). Here, raising the Laplacian prior avoids the fast erosion of relevant parameters over the period following training because it only allows small weights to decay. This better matches the limited loss of association seen between days in the human data of the perceptual expertise task. Raising the Laplacian prior thus results in a biologically plausible form of Bayesian feature selection that is effective in biologically relevant contexts.

Calcium-dependent calcium decay explains STDP in a dynamic model of hippocampal synapses

It is widely accepted that the direction and magnitude of synaptic plasticity depends on post-synaptic calcium flux, where high levels of calcium lead to long-term potentiation and moderate levels lead to long-term depression. At synapses onto neurons in region CA1 of the hippocampus (and many other synapses), NMDA receptors provide the relevant source of calcium. In this regard, post-synaptic calcium captures the coincidence of pre- and post-synaptic activity, due to the blockage of these receptors at low voltage. Previous studies show that under spike timing dependent plasticity (STDP) protocols, potentiation at CA1 synapses requires post-synaptic bursting and an inter-pairing frequency in the range of the hippocampal theta rhythm. We hypothesize that these requirements reflect the saturation of the mechanisms of calcium extrusion from the post-synaptic spine. We test this hypothesis with a minimal model of NMDA receptor-dependent plasticity, simulating slow extrusion with a calcium-dependent calcium time constant. In simulations of STDP experiments, the model accounts for latency-dependent depression with either post-synaptic bursting or theta-frequency pairing (or neither) and accounts for latency-dependent potentiation when both of these requirements are met. The model makes testable predictions for STDP experiments and our simple implementation is tractable at the network level, demonstrating associative learning in a biophysical network model with realistic synaptic dynamics.

Are binary synapses superior to graded weight representations in stochastic attractor networks?

Synaptic plasticity is an underlying mechanism of learning and memory in neural systems, but it is controversial whether synaptic efficacy is modulated in a graded or binary manner. It has been argued that binary synaptic weights would be less susceptible to noise than graded weights, which has impelled some theoretical neuroscientists to shift from the use of graded to binary weights in their models. We compare retrieval performance of models using both binary and graded weight representations through numerical simulations of stochastic attractor networks. We also investigate stochastic attractor models using multiple discrete levels of weight states, and then investigate the optimal threshold for dilution of binary weight representations. Our results show that a binary weight representation is not less susceptible to noise than a graded weight representation in stochastic attractor models, and we find that the load capacities with an increasing number of weight states rapidly reach the load capacity with graded weights. The optimal threshold for dilution of binary weight representations under stochastic conditions occurs when approximately 50% of the smallest weights are set to zero.

Computational consequences of experimentally derived spike-time and weight dependent plasticity rules

We present two weight- and spike-time dependent synaptic plasticity rules consistent with the physiological data of Bi and Poo (J Neurosci 18:10464-10472, 1998). One rule assumes synaptic saturation, while the other is scale free. We extend previous analyses of the asymptotic consequences of weight-dependent STDP to the case of strongly correlated pre- and post-synaptic spiking, more closely resembling associative learning. We further provide a general formula for the contribution of any number of spikes to synaptic drift. Asymptotic weights are shown to principally depend on the correlation and rate of pre- and post-synaptic activity, decreasing with increasing rate under correlated activity, and increasing with rate under uncorrelated activity. Spike train statistics reveal a quantitative effect only in the pre-asymptotic regime, and we provide a new interpretation of the relation between BCM and STDP data.