Biology subjects

Friedman, S. F.

Publications and source records attributed to Friedman, S. F..

5 recordsLinked to original sources

bioRxiv2025

The Latentverse: An Open-Source Benchmarking Toolkit for Evaluating Latent Representations

Self-supervised representation learning is a powerful approach for extracting meaningful features without relying on large amounts of labeled data, making it particularly valuable in fields like healthcare. This enables pretrained models to be shared and fine-tuned with minimal data for various downstream applications. However, evaluating the quality and behavior of these representations remains challenging. To address this, we introduce Latentverse, an open-source library and web-based platform for evaluating latent representations. Latentverse generates detailed reports with visualizations and metrics that provide a comprehensive perspective on different properties of representations, such as clustering, disentanglement, generalization, expressiveness, and robustness. It also allows for the comparison of different representations, enabling developers to refine model architectures and helping users assess how well an embedding model aligns with the requirements of their specific applications. Data and Code AvailabilityThe Latentverse code is available at: https://github.com/broadinstitute/ml4h-latentverse. Institutional Review Board (IRB)This work doesnt require IRB approval.

bioengineering↗

bioRxiv2023

Genetic Architectures of Medical Images Revealed by Registration and Fusion of Multiple Modalities

The advent of biobanks with vast quantities of medical imaging and paired genetic measurements creates huge opportunities for a new generation of genotype-phenotype association studies. However, disentangling biological signals from the many sources of bias and artifacts remains difficult. Using diverse types of medical imaging (i.e. MRIs, ECGs and DXAs), we develop registered and cross-modal generative models. In all cases, we show how registration, both spatial and temporal, guided by domain knowledge or learned de novo, uncovers rich biological information. Remarkably, our findings demonstrate that even extremely lossy transformations, such as registering images onto a single 1D curve (e.g. a circle), can yield robust signals. Conversely, we demonstrate that increasing data dimensionality by integrating multiple modalities can also result in richer representations. Through genome- and phenome-wide association studies (GWAS and PheWAS) of learned embeddings, we uncover significantly more associations with registered and fused modalities than with equivalently trained and sized representations learned from native coordinate spaces. Our findings systematically reveal the crucial role registration plays in enhancing the characterization of physiological states across a broad range of medical imaging data types.

genetics↗

bioRxiv2022

Language Models Learn Sentiment and Substance from 11,000 Psychoactive Experiences

With novel hallucinogens poised to enter psychiatry, we lack a unified framework for quantifying which changes in consciousness are optimal for treatment. Using transformers (i.e. BERT) and 11,816 publicly-available drug testimonials, we first predicted 28-dimensions of sentiment across each narrative, validated with psychiatrist annotations. Secondly, BERT was trained to predict biochemical and demographic information from testimonials. Thirdly, canonical correlation analysis (CCA) linked 52 drugs receptor affinities with testimonial word usage, revealing 11 latent receptor-experience factors, mapped to a 3D cortical atlas. Together, these 3 machine learning methods elucidate a neurobiologically-informed, temporally-sensitive portrait of drug-induced subjective experiences. Different models results converged, revealing a pervasive distinction between lucid and mundane phenomena. MDMA was linked to "Love", DMT and 5-MeO-DMT to "Mystical Experiences", and other tryptamines to "Surprise", "Curiosity" and "Realization". Applying these models to real-time biofeedback, practitioners could harness them to guide the course of therapeutic sessions.

neuroscience↗

bioRxiv2022

Cross-Modal Autoencoder Framework Learns Holistic Representations of Cardiovascular State

A fundamental challenge in diagnostics is integrating multiple modalities to develop a joint characterization of physiological state. Using the heart as a model system, we develop a cross-modal autoencoder framework for integrating distinct data modalities and constructing a holistic representation of cardio-vascular state. In particular, we use our framework to construct such cross-modal representations from cardiac magnetic resonance images (MRIs), containing structural information, and electrocardiograms (ECGs), containing myoelectric information. We leverage the learned cross-modal representation to (1) improve phenotype prediction from a single, accessible phenotype such as ECGs; (2) enable imputation of hard-to-acquire cardiac MRIs from easy-to-acquire ECGs; and (3) develop a framework for performing genome-wide association studies in an unsupervised manner. Our results provide a framework for integrating distinct diagnostic modalities into a common representation that better characterizes physiologic state.

bioinformatics↗

bioRxiv2021

Trips and Neurotransmitters: Discovering Principled Patterns across 6,850 Hallucinogenic Experiences

Psychedelics are thought to alter states of consciousness by disrupting how the higher association cortex governs bottom-up sensory signals. Individual hallucinogenic drugs are usually studied in participants in controlled laboratory settings. Here, we have explored word usage in 6,850 free-form testimonials with 27 drugs through the prism of 40 neurotransmitter receptor subtypes, which were then mapped to 3D coordinates in the brain via their gene transcription levels from invasive tissue probes. Despite the variable subjective nature of hallucinogenic experiences, our pattern-learning approach delineated how drug-induced changes of conscious awareness (e.g., dissolving self-world boundaries or fractal distortion of visual perception) are linked to cortex-wide anatomical distributions of receptor density proxies. The dominant explanatory factor related ego-dissolution-like phenomena to a constellation of 5-HT2A, D2, KOR, and NMDA receptors, anchored especially in the brains deep hierarchy (epitomized by the associative higher-order cortex) and shallow hierarchy (epitomized by the visual cortex). Additional factors captured psychological phenomena in which emotions (5-HT2A and Imidazoline1) were in tension with auditory (SERT, 5-HT1A) or visual (5-HT2A) sensations. Each discovered receptor-experience factor spanned between a higher-level association pole and a sensory input pole, which may relate to the previously reported collapse of hierarchical order among large-scale networks. Simultaneously considering many psychoactive molecules and thousands of natural language descriptions of drug experiences our framework finds the underlying semantic structure and maps it directly to the brain. These advances could assist in unlocking their wide-ranging potential for medical treatment.

neuroscience↗