Biology subjects

Herrero, J.

Publications and source records attributed to Herrero, J..

17 recordsLinked to original sources

bioRxiv2026

Global Neural Oscillations Underlie Performance Variability and Attentional State Fluctuations in Humans

Fluctuations in attentional states, such as mind-wandering (MW), are associated with critical variability in task performance. While fMRI studies highlight the opposing roles of task-positive (e.g., dorsal attention network) and task-negative (e.g., default mode network) systems, the electrophysiological mechanisms underlying these dynamics remain poorly understood. Using intracranial electrocorticography in humans performing a sustained attention task, we identified global oscillatory dynamics linked to attentional shifts. MW was characterized by (i) reduced theta ({theta}) and alpha ({square}) power, (ii) decreased aperiodic signal components, indicating a shift toward cortical inhibition, (iii) enhanced phase synchronization across networks, and (iv) strengthened {theta} phase-behavior correlations ({rho}). These features support a non-network-specific framework in which low-frequency {theta} dynamics--captured by both {theta} power and {rho}--are associated with attentional fluctuations, while aperiodic offset relates to attentional state indirectly through its association with {rho} (Structural Equation Modeling: power[->]state {beta} = -0.118, p = 0.002; {rho}[->]state {beta} = 0.246, p < 0.001; offset[->]{rho} {beta} = -0.222, p < 0.001). Our study provides a unified neurophysiological framework for understanding how spontaneous neural activity can drive attentional fluctuations and performance variability, with implications for research on attention, learning, and neuropsychiatric disorders.

neuroscience↗

bioRxiv2025

A mouse model of PTEN Hamartoma Tumour Syndrome reveals that loss of the nuclear function of PTEN drives macrocephaly, lymphoid overgrowth, and late-onset cancer

PTEN Hamartoma Tumour Syndrome (PHTS) is a rare disorder characterized by germline heterozygous mutations in the PTEN tumour suppressor gene, leading to multi-organ/tissue overgrowth, autism spectrum disorder and increased cancer risk. PHTS individuals display heterogeneity in phenotypes, which has been linked in part to the diverse genetic alterations in the PTEN gene and the multifaceted functions of this protein. Indeed, while PTEN primarily functions as a PIP3 lipid phosphatase in the cytosol, regulating PI3K/AKT signalling, a pathway commonly deregulated in cancer, it also plays crucial roles in maintaining chromosomal stability through nuclear activities such as double strand (ds) DNA damage repair. Recent studies have identified a subset of missense PHTS variants that cause nuclear exclusion of PTEN, impairing its nuclear functions. Here, we present our findings from one such pathogenic variant, PTEN-R173C, frequently found in PHTS and somatic cancers. Using cell biological and mouse modelling approaches, we show that PTEN-R173C has higher PIP3 phosphatase activity than wild-type PTEN, resulting in effective regulation of canonical PI3K/AKT signalling. However, PTEN-R173C is unstable and excluded from the nucleus. Aligning with their near normal PI3K/AKT signalling, Pten+/R173Cmice display a low incidence of solid tumours compared to Pten+/-mice. Pten+/R173C mice also exhibit lymphoid hyperplasia and macrocephaly which correlates with compromised nuclear functions of PTEN-R173C. That nuclear functions are compromised is demonstrated by reduced dsDNA damage repair in Pten+/R173Cmice. Integrating PHTS patient data with findings from our mouse model, our study indicates that nuclear dysfunction of pathogenic PTEN variants is a key factor in predicting the onset of the different PHTS-associated phenotypes. We speculate that late-onset cancer in individuals with nuclear-excluded PTEN results from genetic alterations unrelated to PTEN itself, facilitated by impaired PTEN-mediated dsDNA damage repair.

genetics↗

bioRxiv2025

GIMMEcpg: Global Imputation of Mean CpG MEthylation in Real-time

Whole-genome DNA methylation (methylome) analysis is of broad interest to biomedical research due to its central role in human development and disease. However, generating high-quality methylomes at scale remains challenging due to inherent technical limitations. While imputation has the potential to help overcome this problem, no existing approach adequately addresses the scaling issue. Here, we present GIMMEcpg (Global Imputation of Mean cpg MEthylation), a novel imputation tool that scales efficiently from single samples to large cohort studies. GIMMEcpg uses a custom feature dataset built from known CpG sites within the same dataset to impute missing values by calculating the distance-weighted mean of the methylation value of the two immediately neighbouring CpG sites. We benchmarked GIMMEcpg for speed and accuracy against multiple imputation methods using downsampled datasets produced from high-quality ([~]100x) Whole Genome Bisulfite Sequencing (WGBS) data. With a 10x downsampled dataset, GIMMEcpg was able to process the dataset and impute 9.14 Million CpG sites within 7 seconds (R: 0.78, MAE: +5.6%, RMSE: +10.9%). Our results demonstrate that GIMMEcpg is 39-2,562 times faster than three existing methylation imputation tools (BoostMe, DeepCpG, and MethImpute) while maintaining comparable accuracy. To quantify GIMMEcpgs scalability, we applied it to the most extensive single collection of WGBS data (N=645 at variable coverage) from the EpiATLAS generated by the International Human Epigenome Consortium (IHEC). Using a single, standard CPU server, GIMMEcpg processed and imputed an additional 2.4 billion CpG methylation values across the 645 datasets in less than a day, enriching the EpiATLAS methylome resource by 20%. This demonstrates that GIMMEcpg scales to large cohort studies with only a subtle impact on accuracy, as illustrated by our benchmark. We also developed a machine learning variant, GIMMEcpg.ml, which delivers a higher accuracy compared to existing methodologies. Using the same 10x downsampled benchmarking dataset, GIMMEcpg.ml achieved a Person Correlation of 0.87 compared to the ground truth, representing an improvement of 0.11 over the best performing alternative method. Additionally, GIMMEcpg.ml has a Mean Absolute Error (MAE) of 8.67%, which is 2.63% lower than the most accurate performing alternative. While this enhanced accuracy comes at the cost of increased computation requirements, GIMMEcpg.ml is a useful tool where higher accuracy is preferred over scalability. For increased accessibility, GIMMEcpg is freely available under an MIT license as R and Python packages at https://github.com/ucl-medical-genomics/gimmecpg-r and https://github.com/ucl-medical-genomics/gimmecpg-python, respectively.

bioinformatics↗

bioRxiv2024

Epigenetic investigation of multifocal small intestinal neuroendocrine tumours reveals accelerated ageing of tumours and epigenetic alteration of metabolic genes

BackgroundSmall intestinal neuroendocrine tumours (SI-NETs) are the most common malignancy of the small intestine and around 50% of patients present in clinic with multifocal disease. Recent investigations into the genomic architecture of multifocal SI-NETs have found evidence that these synchronous primary tumours evolve independently of each other. They also have extremely low mutational burden and few known driver genes, suggesting that epigenetic dysregulation may be driving tumorigenesis. Very little is known about epigenetic gene regulation, metabolism and ageing in these tumours, and how these traits differ across multiple tumours within individual patients. MethodsIn this study, we performed the first investigation of genome-wide DNA methylation in multifocal SI-NETs, assessing multiple primary tumours within each patient (n=79 primary tumours from 14 patients) alongside matched metastatic tumours (n=12) and normal intestinal epithelial tissue (n=9). We assessed multifocal SI-NET differential methylation using a novel method, comparing primary tumours with matched normal epithelial tissue and an enterochromaffin-enriched cell line to enrich for tumour-specific effects. This method reduced the identification of false positive methylation differences driven by cell composition differences between tumour and normal epithelial tissue. We also assessed tumour ageing using epigenetic clocks and applied metabolic predictors in the dataset to assess methylation variation across key metabolic genes. ResultsWe have identified 12,392 tumour-specific differentially methylated positions (Bonferroni corrected p<0.05) which were enriched for neural pathways. The expression levels of the genes associated with top sites were also found to be significantly altered in SI-NETs. Age acceleration was observed across SI-NETs and a variability in epigenetic age of tumours within each patient, which we believe is reflecting the order in which tumours have developed. This is supported by the correlation of age acceleration with somatic mutational count in the tumours. We have identified SI-NET associated alterations to the methylation patterns in key metabolic genes compared to matched normal tissue, which is more pronounced in metastatic tumours and tumours harbouring chromosome 18 loss of heterozygosity, indicating metabolic differences in these tumour subtypes. ConclusionsWe have identified accelerated ageing and changes to regulation of metabolic genes, alongside an epigenetic signature of multifocal SI-NETs. These findings add to our understanding of multifocal SI-NET biology and their molecular differences which may be instrumental in the development of these elusive tumours.

cancer biology↗

bioRxiv2023

The neural activity of auditory conscious perception

Although recent work has made significant headway in understanding the temporal and spatial dynamics of the neural mechanisms of conscious perception, much of that work has focused on visual paradigms. To determine whether there are shared mechanisms for perceptual consciousness across sensory modalities, here we developed a task to test within the auditory domain. Participants (n=31) completed an auditory perceptual threshold task while undergoing intracranial electroencephalography (icEEG) for intractable epilepsy. Intracranial recordings from over 2,800 grey matter electrodes representing widespread cortical coverage were analyzed for power in the high gamma range (40-115 Hz)--a frequency range that reflects local neural activity. For trials that were perceived, we find activity in early auditory regions which is accompanied by activity in the right caudal middle frontal gyrus, and shortly thereafter by activity in non-auditory thalamus. This is followed by a wave of activity that sweeps through the higher auditory association regions and into parietal and frontal cortices, similar to the wave observed in our visual conscious perception paradigm. However, for not perceived trials, we find that significant activity is restricted to early auditory regions (and areas immediately adjacent to the Sylvian fissure). These findings show that the broad anatomical regions of cortical and subcortical networks involved in auditory perception are similar to the networks observed with vision, suggesting shared general mechanisms for conscious perception.

neuroscience↗

bioRxiv2022

Co-option of endogenous retroviruses through genetic escape from TRIM28 repression

Endogenous retroviruses (ERVs) have rewired host gene networks through co-option of their enhancers. To explore which ERVs get co-opted and why, we tracked the epigenetic fate of murine IAPEz elements using an in vitro model of embryonic stem cell (ESC) to neural progenitor cell (NPC) differentiation. TRIM28-repression depended on a 190bp sequence, previously shown to confer IAPEzs with retrotransposition activity. A subset of IAPEzs ([~]15%) exhibit genetic divergence from this sequence, which we term escapees. While repressed IAPEzs succumb to a previously undocumented epigenetic handover from H3K9me3 in ESCs to H3K27me3 in NPCs, escapee IAPEzs evade repression, resulting in their transcriptional derepression in NPCs. Escapee IAPEzs enhance expression of nearby neural genes and contribute to gene expression differences between mouse strains, which we discern by employing IAPEz insertion polymorphisms. In sum, co-opted ERVs stem from genetic escapees that have lost vital sequences required for both TRIM28 restriction and autonomous retrotransposition. HIGHLIGHTSO_LITracking the epigenetic fate of ERVs through neural differentiation reveals which ERVs are subject to co-option and why. C_LIO_LIMost ERVs succumb to H3K9me3 deposition in ESCs with H3K27me3 memory in NPCs. C_LIO_LIERVs that act as enhancers to nearby neural genes have undergone genetic escape from TRIM28 repression. C_LIO_LIEpigenetic evasion is an evolutionary trade-off that comes with loss of sequences necessary for retrotransposition. C_LI O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=130 SRC="FIGDIR/small/497016v1_ufig1.gif" ALT="Figure 1"> View larger version (21K): org.highwire.dtl.DTLVardef@175b6c0org.highwire.dtl.DTLVardef@1313692org.highwire.dtl.DTLVardef@1f5e33forg.highwire.dtl.DTLVardef@1a5ad3_HPS_FORMAT_FIGEXP M_FIG O_FLOATNOGraphical abstract.C_FLOATNO C_FIG

molecular biology↗

bioRxiv2022

Complex genotype-phenotype relationships shape the response to treatment of Down Syndrome Childhood Acute Lymphoblastic Leukaemia

Extensive genetic and epigenetic variegation has been demonstrated in many malignancies. Importantly, their interplay has the potential to contribute to disease progression and treatment resistance. To shed light on the complex relationships between these different sources of intra-tumour heterogeneity, we explored their relative contributions to the evolutionary dynamics of Acute Lymphoblastic Leukemia (ALL) in children with Down syndrome, which has particularly poor prognosis. We quantified the tumour propagating potential of genetically distinct sub-clones using serial transplantation assays and SNP-arrays. While most leukaemias were characterized by a single dominant subclone, others were highly heterogeneous. Importantly, we provide clear and direct evidence that genotypes and phenotypes with functional relevance to leukemic progression and treatment resistance can co-segregate within the disease. Hence, individual genetic lesions can be restricted to well-defined cell immunophenotypes, corresponding to different stages of the leukemic differentiation hierarchy and varied proliferation potentials. As a result of this difference in fitness, which can be accurately quantified via competitive transplantation assays, matching diagnostic, post-treatment, and relapse leukemias can be dominated by different genotypes, including pre-leukemic clones persisting throughout the disease progression and treatment. Intriguingly, plasticity also appears to be a temporally defined property that can segregate with genotype. These results suggest that Down Syndrome ALL should be viewed as a complex matrix of cells exhibiting genetic and epigenetic heterogeneity that foster extensive clonal evolution and competition. Therapeutic intervention reshapes this eco-system and may provide the right conditions for the preferential expansion of selected compartments and subsequently relapse.

cancer biology↗

bioRxiv2021

Aberrant paracrine signalling for bone remodelling underlies the mutant histone-driven giant cell tumour of bone.

Oncohistones represent compelling evidence for a causative role of epigenetic perturbations in cancer. Giant cell tumours of bone (GCTs) are characterised by a mutated histone H3.3 as the sole genetic driver present in bone-forming osteoprogenitor cells but absent from abnormally large bone-resorbing osteoclasts which represent the hallmark of these neoplasms. While these striking features imply a pathogenic interaction between mesenchymal and myelomonocytic lineages during GCT development, the underlying mechanisms remain unknown. We show that the changes in the transcriptome and epigenome in the mesenchymal cells caused by the H3.3-G34W mutation contribute to increase osteoclast recruitment in part via reduced expression of the TGF{beta}-like soluble factor, SCUBE3. In turn, osteoclasts secrete unregulated amounts of SEMA4D enhancing proliferation of mutated osteoprogenitors and arresting their maturation. These findings provide a mechanism by which GCTs undergo differentiation upon denosumab treatment, a drug that depletes osteoclasts. In contrast, gain of hTERT activity, commonly found in malignant GCT, makes neoplastic cells insensitive to osteoclasts, predicting the unresponsiveness to denosumab. We provide a mechanism for GCT initiation and its response to current treatment, the basis of which is dysfunctional cross-talk between bone-forming and bone-resorbing cells, emphasising the importance of tumor/microenvironment bidirectional interactions in tumorigenesis.

cancer biology↗

bioRxiv2021

The Th1 cell regulatory circuitry is largely conserved between human and mouse

Gene expression programmes controlled by lineage-determining transcription factors are often conserved between species. However, infectious diseases have exerted profound evolutionary pressure, and therefore the genes regulated by immune-specific transcription factors might be expected to exhibit greater divergence due to exposure to species-specific pathogens. T-bet (Tbx21) is the immune-specific lineage-defining transcription factor for T helper type I (Th1) immunity, which is fundamental for the immune response to intracellular pathogens but also underlies inflammatory diseases. We therefore compared T-bet genomic targets between mouse and human CD4+ T cells and correlated T-bet binding patterns with species-specific gene expression. Remarkably, we show that the vast majority of T-bet regulated genes are conserved between mouse and human, either via preservation of a binding site or via an alternative binding site associated with transposon-linked insertion. We also identified genes that are specifically targeted by T-bet in humans or mice and which exhibited species-specific expression. These results provide a genome-wide cross-species comparison of T-bet target gene regulation that will enable more accurate translation of genetic targets and therapeutics from pre-clinical models of inflammatory disease into human clinical trials.

immunology↗

bioRxiv2020

Functional characterization of human Heschl's gyrus in response to natural speech

Heschls gyrus (HG) is a brain area that includes the primary auditory cortex in humans. Due to the limitations in obtaining direct neural measurements from this region during naturalistic speech listening, the functional organization and the role of HG in speech perception remains uncertain. Here, we used intracranial EEG to directly record the neural activity in HG in eight neurosurgical patients as they listened to continuous speech stories. We studied the spatial distribution of acoustic tuning and the organization of linguistic feature encoding. We found a main gradient of change from posteromedial to anterolateral parts of HG. Along this direction, we observed a decrease in frequency and temporal modulation tuning, and an increase in phonemic representation, speaker normalization, speech-sensitivity, and response latency. We did not observe a difference between the two brain hemispheres. These findings reveal a functional role for HG in processing and transforming simple to complex acoustic features and informs neurophysiological models of speech processing in the human auditory cortex.

neuroscience↗

bioRxiv2020

GenomeChronicler: The Personal Genome Project UK Genomic Report Generator Pipeline

In recent years, there has been a significant increase in whole genome sequencing data of individual genomes produced by research projects as well as direct to consumer service providers. While many of these sources provide their users with an interpretation of the data, there is a lack of free, open tools for generating reports exploring the data in an easy to understand manner. GenomeChronicler was developed as part of the Personal Genome Project UK (PGP-UK) to address this need. PGP-UK provides genomic, transcriptomic, epigenomic and self-reported phenotypic data under an open-access model with full ethical approval. As a result, the reports generated by GenomeChronicler are intended for research purposes only and include information relating to potentially beneficial and potentially harmful variants, but without clinical curation. GenomeChronicler can be used with data from whole genome or whole exome sequencing, producing a genome report containing information on variant statistics, ancestry and known associated phenotypic traits. Example reports are available from the PGP-UK data page (personalgenomes.org.uk/data). The objective of this method is to leverage existing resources to find known phenotypes associated with the genotypes detected in each sample. The provided trait data is based primarily upon information available in SNPedia, but also collates data from ClinVar, GETevidence and gnomAD to provide additional details on potential health implications, presence of genotype in other PGP participants and population frequency of each genotype. The analysis can be run in a self-contained environment without requiring internet access, making it a good choice for cases where privacy is essential or desired: any third party project can embed GenomeChronicler within their off-line safe-haven environments. GenomeChronicler can be run for one sample at a time, or in parallel making use of the Nextflow workflow manager. The source code is available from GitHub (https://github.com/PGP-UK/GenomeChronicler), container recipes are available for Docker and Singularity, as well as a pre-built container from SingularityHub (https://singularity-hub.org/collections/3664) enabling easy deployment in a variety of settings. Users without access to computational resources to run GenomeChronicler can access the software from the Lifebit CloudOS platform (https://lifebit.ai/cloudos) enabling the production of reports and variant calls from raw sequencing data in a scalable fashion.

bioinformatics↗

bioRxiv2019

Escape from nonsense mediated decay associates with anti-tumor immunogenicity

Frameshift insertion/deletions (fs-indels) are an infrequent but potentially highly immunogenic mutation subtype. Although fs-indel transcripts are susceptible to degradation through the non-sense mediated decay (NMD) pathway, we hypothesise that some fs-indels escape degradation and lead to an increased abundance of tumor specific neoantigens, that are highly distinct from self. We analysed matched DNA and RNA sequencing data from TCGA, and five separate melanoma cohorts treated with immunotherapy. Using allele-specific expression analysis we show that expressed fs-indels were enriched in genomic positions predicted to escape NMD, and associated with higher protein expression, consistent with degradation escape ("NMD-escape"). Across four independent cohorts, fs-indel NMD-escape mutations were found to be significantly associated with clinical benefit to checkpoint inhibitor (CPI) therapy (Pmeta=0.0039), a stronger association than either nsSNV (Pmeta=0.073) or fs-indel (Pmeta=0.064) count. NMD-escape mutations were additionally shown to have independent predictive power in the "low-TMB" setting, and may serve as a biomarker to rescue patients judged ineligible for CPI based on overall TMB, but still with a high chance of response (low-TMB cohort: NMD-escape-positive % clinical benefit=53%, NMD-escape-negative % clinical benefit=16%, P=0.0098). Furthermore, in an adoptive cell therapy (ACT) treated cohort, NMD-escape mutation count was the most significant biomarker associated with clinical benefit (P=0.021). Analysis of functional T-cell reactivity screens from recent personalized vaccine and CPI studies shows direct evidence of fs-indel derived neoantigens eliciting patient anti-tumor immune response (n=15). We additionally observe a subset of fs-indel mutations, with highly elongated neo open reading frames, which are found to be significantly enriched for immunogenic reactivity in these patient studies (P=0.0032). Finally, consistent with the potency of NMD-escape derived neo-antigens and ongoing immune-editing, NMD-escape fs-indels appear to be under negative selective pressure in untreated TCGA cases. Given the strongly immunogenic potential, and relatively rare nature of NMD-escape fs-indels, these alterations may be attractive candidates in immunotherapy biomarker optimisation and neoantigen ACT or vaccine strategies.

cancer biology↗

bioRxiv2019

Estimating and interpreting nonlinear receptive fields of sensory responses with deep neural network models

Sensory processing by neural circuits includes numerous nonlinear transformations that are critical to perception. Our understanding of these nonlinear mechanisms, however, is hindered by the lack of a comprehensive and interpretable computational framework that can model and explain nonlinear signal transformations. Here, we propose a data-driven framework based on deep neural network regression models that can directly learn any nonlinear stimulus-response mapping. A key component of this approach is an analysis method that reformulates the exact function of the trained neural network as a collection of stimulus-dependent linear functions. This locally linear receptive field interpretation of the network function enables straightforward comparison with conventional receptive field models and uncovers nonlinear encoding properties. We demonstrate the efficacy of this framework by predicting the neural responses recorded invasively from the auditory cortex of neurosurgical patients as they listened to speech. Our method significantly improves the prediction accuracy of auditory cortical responses particularly in nonprimary areas. Moreover, interpreting the functions learned by neural networks uncovered three distinct types of nonlinear transformations of speech that varied considerably in primary and nonprimary auditory regions. By combining two desired properties of a computational sensory-response model; the ability to capture arbitrary stimulus-response mappings and maintaining model interpretability, this data-driven method can lead to better neurophysiological models of the sensory processing.

neuroscience↗

bioRxiv2019

Cortical encoding of melodic expectations in human temporal cortex

Humans engagement in music rests on underlying elements such as the listeners cultural background and general interest in music, all shaping the way music is processed in the brain and perceived. Crucially, these factors modulate how listeners anticipate musical events, a process inducing instantaneous neural responses as the music confronts these expectations. Measuring such neural correlates would represent a direct window into high-level brain processing of music. Here we recorded electroencephalographic and electrocorticographic brain responses as participants listened to Bach melodies. We assessed the relative contributions of the acoustic versus melodic components of the music to the neural signal. Acoustic features included envelope and its derivative. Melodic features included information on melodic progressions (pitch) and their tempo (onsets), which were extracted from a Markov model predicting the next note based on a corpus of Western music and the preceding proximal musical context. We related the music to brain activity with a linear temporal response function, and demonstrated that cortical responses to music encode melodic expectations. Specifically, individual-subject neural signals were better predicted by a combination of acoustic and melodic expectation features than by either alone. This effect was most pronounced at response latencies up to 350ms, and in both planum temporale and Heschls gyrus. Finally, expectations of pitch and onset-time of musical notes exerted independent cortical effects, and such influences were modulated by the listeners musical expertise. Overall, this study demonstrates how the interplay of experimental and theoretical approaches can yield novel insights into the cortical encoding of melodic expectations.

neuroscience↗

bioRxiv2019

The Personal Genome Project-UK: an open access resource of human multi-omics data

Integrative analysis of multi-omics data is a powerful approach for gaining functional insights into biological and medical processes. Conducting these multifaceted analyses on human samples is often complicated by the fact that the raw sequencing output is rarely available under open access. The Personal Genome Project UK (PGP-UK) is one of few resources that recruits its participants under open consent and makes the resulting multi-omics data freely and openly available. As part of this resource, we describe the PGP-UK multi-omics reference panel consisting of ten genomic, methylomic and transcriptomic data. Specifically, we outline the data processing, quality control and validation procedures which were implemented to ensure data integrity and exclude sample mix-ups. In addition, we provide a REST API to facilitate the download of the entire PGP-UK dataset. The data are also available from two cloud-based environments, providing platforms for free integrated analysis. In conclusion, the genotype-validated PGP-UK multi-omics human reference panel described here provides a valuable new open access resource for integrated analyses in support of personal and medical genomics.

bioinformatics↗

bioRxiv2018

Reconstructing intelligible speech from the human auditory cortex

Auditory stimulus reconstruction is a technique that finds the best approximation of the acoustic stimulus from the population of evoked neural activity. Reconstructing speech from the human auditory cortex creates the possibility of a speech neuroprosthetic to establish a direct communication with the brain and has been shown to be possible in both overt and covert conditions. However, the low quality of the reconstructed speech has severely limited the utility of this method for brain-computer interface (BCI) applications. To advance the state-of-the-art in speech neuroprosthesis, we combined the recent advances in deep learning with the latest innovations in speech synthesis technologies to reconstruct closed-set intelligible speech from the human auditory cortex. We investigated the dependence of reconstruction accuracy on linear and nonlinear (deep neural network) regression methods and the acoustic representation that is used as the target of reconstruction, including auditory spectrogram and speech synthesis parameters. In addition, we compared the reconstruction accuracy from low and high neural frequency ranges. Our results show that a deep neural network model that directly estimates the parameters of a speech synthesizer from all neural frequencies achieves the highest subjective and objective scores on a digit recognition task, improving the intelligibility by 65% over the baseline method which used linear regression to reconstruct the auditory spectrogram. These results demonstrate the efficacy of deep learning and speech synthesis algorithms for designing the next generation of speech BCI systems, which not only can restore communications for paralyzed patients but also have the potential to transform human-computer interaction technologies.

neuroscience↗

bioRxiv2018

PGP-UK: a research and citizen science hybrid project in support of personalized medicine

Molecular analyses such as whole-genome sequencing have become routine and are expected to be transformational for future healthcare and lifestyle decisions. Population-wide implementation of such analyses is, however, not without challenges, and multiple studies are ongoing to identify what these are and explore how they can be addressed. Defined as a research project, the Personal Genome Project UK (PGP-UK) is part of the global PGP network and focuses on open data sharing and citizen science to advance and accelerate personalized genomics and medicine. Here we report our findings on using an open consent recruitment protocol, active participant involvement, open access release of personal genome, methylome and transcriptome data and associated analyses, including 47 new variants predicted to affect gene function and innovative reports based on the analysis of genetic and epigenetic variants. For this pilot study, we recruited ten participants willing to actively engage as citizen scientists with the project. In addition, we introduce Genome Donation as a novel mechanism for openly sharing previously restricted data and discuss the first three donations received. Lastly, we present GenoME, a free, open-source educational app suitable for the lay public to allow exploration of personal genomes. Our findings demonstrate that citizen science-based approaches like PGP-UK have an important role to play in the public awareness, acceptance and implementation of genomics and personalized medicine.

genomics↗