Biology subjects

Vogelstein, J.

Publications and source records attributed to Vogelstein, J..

12 recordsLinked to original sources

bioRxiv2025

Statistically valid explainable black-box machine learning: applications in sex classification across species using brain imaging

Sex classification using neuroimaging data has the potential to revolutionize personalized diagnostics by revealing subtle structural brain differences that underlie sex-specific disease risks. Despite the promise of machine learning, traditional methods often fall short in providing both high classification accuracy and interpretable, statistically validated feature importance scores for high-dimensional imaging data. This gap is particularly evident when conventional techniques such as random forests, LIME, and SHAP are applied, as they struggle with complex feature interactions and managing noise in large datasets. We address this challenge by developing an integrated framework that combines Oblique Random Forests (ORFs) with a novel, permutation-based feature importance testing algorithm. ORFs extend traditional random forests by employing oblique decision boundaries through linear combinations of features, thereby capturing intricate interactions inherent in neuroimaging data. Our feature importance testing method, NEOFIT, rigorously quantifies the significance of each feature by generating null distributions and corrected p-values. We first validate our approach using simulated datasets, establishing its robustness and scalability under controlled conditions. We then apply our method to classify sex from both voxel-wise structural MRI and cortical thickness data in humans and macaques, facilitating direct cross-species comparisons. ORFs achieves AUC > 0.80 on human data, and > 0.70 on macaque data, while NEOFIT identifies statistically significant features aligned with sex-dimorphic neuroanatomy. Our results demonstrate that the proposed framework not only enhances classification performance but also provides clear, interpretable insights into the neuroanatomical features that distinguish sexes. These methodological advancements pave the way for improved diagnostic tools and contribute to a deeper understanding of the evolutionary basis of sex differences in brain structure.

neuroscience↗

bioRxiv2024

Why do we have so many excitatory neurons?

The emerging electron microscopy connectome datasets provides connectivity maps of the brains at single cell resolution, enabling us to estimate various network statistics, such as connectedness. We desire the ability to assess how the functional complexity of these networks depends on these network statistics. To this end, we developed an analysis pipeline and a statistic, XORness, which quantifies the functional complexity of these networks with varying network statistics. We illustrate that actual connectomes have high XORness, as do generated connectomes with the same network statistics, suggesting a normative role for functional complexity in guiding the evolution of connectomes, and providing clues to guide the development of artificial neural networks.

neuroscience↗

bioRxiv2023

BrainLine: An Open Pipeline for Connectivity Analysis of Heterogeneous Whole-Brain Fluorescence Volumes

Whole-brain fluorescence images require several stages of computational processing to fully reveal the neuron morphology and connectivity information they contain. However, these computational tools are rarely part of an integrated pipeline. Here we present BrainLine, an open-source pipeline that interfaces with existing software to provide registration, axon segmentation, soma detection, visualization and analysis of results. By implementing a feedback based training paradigm with BrainLine, we were able to use a single learning algorithm to accurately process a diverse set of whole-brain images generated by light-sheet microscopy. BrainLine is available as part of our Python package brainlit: http://brainlit.neurodata.io/.

bioengineering↗

bioRxiv2022

Suboptimal phenotypic reliability impedes reproducible human neuroscience

Summary ParagraphBiomarkers of behavior and psychiatric illness for cognitive and clinical neuroscience remain out of reach1-4. Suboptimal reliability of biological measurements, such as functional magnetic resonance imaging (fMRI), is increasingly cited as a primary culprit for discouragingly large sample size requirements and poor reproducibility of brain-based biomarker discovery1,5-7. In response, steps are being taken towards optimizing MRI reliability and increasing sample sizes8-11, though this will not be enough. Optimizing biological measurement reliability and increasing sample sizes are necessary but insufficient steps for biomarker discovery; this focus has overlooked the other side of the equation - the reliability of clinical and cognitive assessments - which are often suboptimal or unassessed. Through a combination of simulation analysis and empirical studies using neuroimaging data, we demonstrate that the joint reliability of both biological and clinical/cognitive phenotypic measurements must be optimized in order to ensure biomarkers are reproducible and accurate. Even with best-case scenario high reliability neuroimaging measurements and large sample sizes, we show that suboptimal reliability of phenotypic data (i.e., clinical diagnosis, behavioral and cognitive measurements) will continue to impede meaningful biomarker discovery for the field. Improving reliability through development of novel assessments of phenotypic variation is needed, but it is not the sole solution. We emphasize the potential to improve the reliability of established phenotypic methods through aggregation across multiple raters and/or measurements12-15, which is becoming increasingly feasible with recent innovations in data acquisition (e.g., web- and smart-phone-based administration, ecological momentary assessment, burst sampling, wearable devices, multimodal recordings)16-20. We demonstrate that such aggregation can achieve better biomarker discovery for a fraction of the cost engendered by large-scale samples. Although the current study has been motivated by ongoing developments in neuroimaging, the prioritization of reliable phenotyping will revolutionize neurobiological and clinical endeavors that are focused on brain and behavior.

neuroscience↗

bioRxiv2021

A low-resource reliable pipeline to democratize multi-modal connectome estimation and analysis

Connectomics--the study of brain networks--provides a unique and valuable opportunity to study the brain. Research in human connectomics, leveraging functional and diffusion Magnetic Resonance Imaging (MRI), is a resource-intensive practice. Typical analysis routines require significant computational capabilities and subject matter expertise. Establishing a pipeline that is low-resource, easy to use, and off-the-shelf (can be applied across multifarious datasets without parameter tuning to reliably estimate plausible connectomes), would significantly lower the barrier to entry into connectomics, thereby democratizing the field by empowering a more diverse and inclusive community of connectomists. We therefore introduce MRI to Graphs (m2g). To illustrate its properties, we used m2g to process MRI data from 35 different studies ({approx} 6,000 scans) from 15 sites without any manual intervention or parameter tuning. Every single scan yielded an estimated connectome that adhered to established properties, such as stronger ipsilateral than contralateral connections in structural connectomes, and stronger homotopic than heterotopic correlations in functional connectomes. Moreover, the connectomes estimated by m2g are more similar within individuals than between them, suggesting that m2g preserves biological variability. m2g is portable, and can run on a single CPU with 16 GB of RAM in less than a couple hours, or be deployed on the cloud using its docker container. All code is available on https://github.com/neurodata/m2g and documentation is available on docs.neurodata.io/m2g.

bioinformatics↗

bioRxiv2020

Modern Machine Learning: Partition & Vote

We present modern machine learning, focusing on the state-of-the-art classification methods of decision forests and deep networks, as partition and vote schemes. This illustrative presentation allows for both a unified basic understanding of how these methods work from the perspective of classical statistical pattern recognition as well as useful basic insight into their relationship with each other ... and potentially with brain functioning.

neuroscience↗

bioRxiv2020

TOWARD A CONNECTIVITY GRADIENT-BASED FRAMEWORK FOR REPRODUCIBLE BIOMARKER DISCOVERY

AO_SCPLOWBSTRACTC_SCPLOWDespite myriad demonstrations of feasibility, the high dimensionality of fMRI data remains a critical barrier to its utility for reproducible biomarker discovery. Recent studies applying dimensionality reduction techniques to resting-state fMRI (R-fMRI) have unveiled neurocognitively meaningful connectivity gradients that are present in both human and primate brains, and appear to differ meaningfully among individuals and clinical populations. Here, we provide a critical assessment of the suitability of connectivity gradients for biomarker discovery. Using the Human Connectome Project (discovery subsample=209; two replication subsamples= 209x2) and the Midnight scan club (n=9), we tested the following key biomarker traits - reliability, reproducibility and predictive validity - of functional gradients. In doing so, we systematically assessed the effects of three analytical settings, including i) dimensionality reduction algorithms (i.e., linear vs. non-linear methods), ii) input data types (i.e., raw time series, [un-]thresholded functional connectivity), and iii) amount of the data (R-fMRI time-series lengths). We found that the reproducibility of functional gradients across algorithms and subsamples is generally higher for those explaining more variances of whole-brain connectivity data, as well as those having higher reliability. Notably, among different analytical settings, a linear dimensionality reduction (principal component analysis in our study), more conservatively thresholded functional connectivity (e.g., 95-97%) and longer time-series data (at least [≥]20mins) was found to be preferential conditions to obtain higher reliability. Those gradients with higher reliability were able to predict unseen phenotypic scores with a higher accuracy, highlighting reliability as a critical prerequisite for validity. Importantly, prediction accuracy with connectivity gradients exceeded that observed with more traditional edge-based connectivity measures, suggesting the added value of a low-dimensional gradient approach. Finally, the present work highlights the importance and benefits of systematically exploring the parameter space for new imaging methods before widespread deployment. HO_SCPLOWIGHLIGHTSC_SCPLOW- There is a growing need to identify benchmark parameters in advancing functional connectivity gradients into a reliable biomarker. - Here, we explored multidimensional parameter space in calculating functional gradients to improve their reproducibility, reliability and predictive validity. - We demonstrated that more reproducible and reliable gradient markers tend to have higher predictive power for unseen phenotypic scores across various cognitive domains. - We showed that the low-dimensional connectivity gradient approach could outperform raw edge-based analyses in terms of predicting phenotypic scores. - We highlight the necessity of optimizing parameters for new imaging methods before their widespread deployment.

neuroscience↗

bioRxiv2020

Visualizing synaptic plasticity in vivo by large-scale imaging of endogenous AMPA receptors

Elucidating how synaptic molecules such as AMPA receptors mediate neuronal communication and tracking their dynamic expression during behavior is crucial to understand cognition and disease, but current technological barriers preclude large-scale exploration of molecular dynamics in vivo. We have developed a suite of innovative methodologies that break through these barriers: a new knockin mouse line with fluorescently tagged endogenous AMPA receptors, two-photon imaging of hundreds of thousands of labeled synapses in behaving mice, and computer-vision-based automatic synapse detection. Using these tools, we can longitudinally track how the strength of synapses changes during behavior. We used this approach to generate an unprecedentedly detailed spatiotemporal map of synaptic plasticity underlying sensory experience. More generally, these tools can be used as an optical probe capable of measuring functional synapse strength across entire brain areas during any behavioral paradigm, describing complex system-wide changes with molecular precision.

neuroscience↗

bioRxiv2019

Deep learning for brains?: Different linear and nonlinear scaling in UK Biobank brain images vs. machine-learning datasets

In recent years, deep learning has unlocked unprecedented success in various domains, especially in image, text, and speech processing. These breakthroughs may hold promise for neuroscience and especially for brain-imaging investigators who start to analyze thousands of participants. However, deep learning is only beneficial if the data have nonlinear relationships and if they are exploitable at currently available sample sizes. We systematically profiled the performance of deep models, kernel models, and linear models as a function of sample size on UK Biobank brain images against established machine learning references. On MNIST and Zalando Fashion, prediction accuracy consistently improved when escalating from linear models to shallow-nonlinear models, and further improved when switching to deep-nonlinear models. The more observations were available for model training, the greater the performance gain we saw. In contrast, using structural or functional brain scans, simple linear models performed on par with more complex, highly parameterized models in age/sex prediction across increasing sample sizes. In fact, linear models kept improving as the sample size approached [~]10,000 participants. Our results indicate that the increase in performance of linear models with additional data does not saturate at the limit of current feasibility. Yet, nonlinearities of common brain scans remain largely inaccessible to both kernel and deep learning methods at any examined scale.

bioinformatics↗

bioRxiv2019

Thermal sensors improve wrist-worn position tracking

Wearable devices provide a means of tracking hand position in relation to the head, but have mostly relied on wrist-worn inertial measurement unit sensors and proximity sensors, which are inadequate for identifying specific locations. This limits their utility for accurate and precise monitoring of behaviors or providing feedback to guide behaviors. A potential clinical application is monitoring body-focused repetitive behaviors (BFRBs), recurrent, injurious behaviors directed toward the body, such as nail biting and hair pulling, that are often misdiagnosed and undertreated. Here, we demonstrate that including thermal sensors achieves higher accuracy in position tracking when compared against inertial measurement unit and proximity sensor data alone. Our Tingle device distinguished between behaviors from six locations on the head across 39 adult participants, with high AUROC values (best was back of the head: median (1.0), median absolute deviation (0.0); worst was on the cheek: median (0.93), median absolute deviation (0.09)). This study presents preliminary evidence of the advantage of including thermal sensors for position tracking and the Tingle wearable devices potential use in a wide variety of settings, including BFRB diagnosis and management.

bioinformatics↗

bioRxiv2018

Improving Corticostriatal Parcellation Through Multilevel Bagging with PyBASC

Increasing the reproducibility of neuroimaging measurement addresses a central impediment to the advancement of human neuroscience and its clinical applications. Recent efforts demonstrating variance in functional brain organization within and between individuals shows a need for improving reproducibility of functional parcellations without long scan times. We apply bootstrap aggregation, or bagging, to the problem of improving reproducibility in functional parcellation. We use two large datasets to demonstrate that compared to a standard clustering framework, bagging improves the reproducibility and test-retest reliability of both cortical and subcortical functional parcellations across a range of sites, scanners, samples, scan lengths, clustering algorithms, and clustering parameters (e.g., number of clusters, spatial constraints). With as little as six minutes of scan time, bagging creates more reproducible group and individual level parcellations than standard approaches with twice as much data. This suggests that regardless of the specific parcellation strategy employed, bagging may be a key method for improving functional parcellation and bringing functional neuroimaging-based measurement closer to clinical impact.

neuroscience↗

bioRxiv2017

A Comprehensive Cloud Framework for Accurate and Reliable Human Connectome Estimation and Meganalysis

Modern scientific discovery depends on collecting large heterogeneous datasets with many sources of variability, and applying domain-specific pipelines from which one can draw insight or clinical utility. For example, macroscale connectomics studies require complex pipelines to process raw functional or diffusion data and estimate connectomes. Individual studies tend to customize pipelines to their needs, raising concerns about their reproducibility, which add to a longer list of factors that may differ across studies and result in failures to replicate (including sampling, experimental design, and data acquisition protocols). Mitigating these issues requires multi-study datasets and the development of pipelines that can be applied across them. We developed NeuroDatas MRI to Graphs (NDMG) pipeline using several functional and diffusion studies, including the Consortium for Reliability and Reproducability, to estimate connectomes. Without any manual intervention or parameter tuning, NDMG ran on 25 different studies ({approx}6,000 scans) from 19 sites, with each scan resulting in a biologically plausible connectome (as assessed by multiple quality assurance metrics at each processing stage). For each study, the connectomes from NDMG are more similar within than across individuals, indicating that NDMG is preserving biological variability. Moreover, the connectomes exhibit near perfect consistency for certain connectional properties across every scan, individual, study, site and modality; these include stronger ipsilateral than contralateral connections and stronger homotopic than heterotopic connections. Yet, the magnitude of the differences varied across individuals and studies--much more so when pooling data across sites, even after controlling for study, site, and basic demographic variables (i.e., age, sex, and ethnicity). This indicates that other experimental variables (possibly those not measured or reported) are contributing to this variability, which if not accounted for can limit the value of aggregate datasets, as well as expectations regarding the accuracy of findings and likelihood of replication. We therefore provide a set of principles to guide the development of pipelines capable of pooling data across studies while maintaining biological variability and minimizing measurement error. This open science approach provides us with an opportunity to understand and eventually mitigate spurious results for both past and future studies.

neuroscience↗