Biology subjects

Berger, J. I.

Publications and source records attributed to Berger, J. I..

16 recordsLinked to original sources

bioRxiv2026

Neural Bases for Working Memory of Natural Sound Textures

Holding sound objects in mind is critical to understanding the sound world around us. Auditory texture perception is thought to rely on extracting statistical properties of sounds over time, raising the question of how encoding duration influences subsequent maintenance of auditory texture information. We investigated this question using behavioural performance and fMRI while participants performed two delayed match-to-sample tasks: a texture discrimination and a comparison loudness discrimination task. In the texture task, participants judged whether two sequentially presented sounds belonged to the same texture type, whereas in the loudness task, participants judged whether the sounds, drawn from the same texture type, had the same loudness. Stimuli were presented for either 0.1s or 2s. Discrimination sensitivity increased with duration in both tasks, but more strongly for texture discrimination. fMRI analysis revealed distinct effects of duration on univariate and multivariate activity in the auditory cortex and medial temporal lobe (MTL) during maintenance. We found no univariate duration effect in the auditory cortex during maintenance, but multivariate patterns encoded texture category more strongly for longer stimuli, regardless of task. In contrast, MTL activity was modulated by sound duration in both tasks, but there was no effect of duration on texture category decoding. There was a task effect in MTL where multivariate patterns in posterior hippocampal and parahippocampal regions distinguished texture from loudness tasks following 2s stimuli. Together, these findings show effects of duration on sound texture representations during maintenance and that MTL representations reflect the behavioural context in which that information is maintained.

neuroscience↗

bioRxiv2026

Ordinal position in auditory sequences is encoded in population dynamics in macaque frontal cortex

The ability to extract structure from sensory sequences requires tracking not only stimulus identity but also ordinal position within a sequence. How and where ordinal information is represented in the primate brain remains unclear. Here, we recorded activity in populations of single neurons simultaneously from auditory cortex and ventrolateral prefrontal cortex in macaques passively listening to structured auditory sequences. Population-level analyses revealed that frontal cortex activity evolved progressively across the sequence, occupying distinct regions of neural state space for successive ordinal positions. In contrast, auditory cortex population activity reset with each element, producing stereotyped, stimulus-locked trajectories. Consistent with this dissociation, frontal cortex populations robustly decoded ordinal position independently of stimulus identity, whereas auditory cortex showed weak positional decoding despite strong identity representations. Frontal neurons further exhibited longer intrinsic timescales and reduced synchronization to element presentation, supporting integration of information across multiple elements. Together, these findings identify frontal activity as a key substrate for ordinal position encoding during auditory sequence processing, while auditory cortex primarily supports element-specific sensory representations.

neuroscience↗

bioRxiv2026

Age-Related Changes in the Neural Dynamics of Speech Production Following Noninvasive Brain Stimulation

IntroductionNon-invasive brain stimulation is a promising technique to restore neural function in older adults by enhancing cortical excitability. In the present study, we examined the behavioral and neural effects of three stimulation protocols over the supplementary motor area on speech production in older adults. These protocols included fixed-frequency and personalized high-definition transcranial alternating current stimulation (HD-tACS) and high-definition transcranial random noise stimulation (HD-tRNS). We also compared the effects of personalized HD-tACS and tRNS between younger and older adults. Two groups, older adults (mean age: 69 y {+/-} 8.18) and younger adults (mean age: 19.87 y {+/-} 1.54), participated in this study. Older adults attended a four-session experiment including sham, fixed-frequency beta-HD-tACS (17 Hz), personalized beta-HD-tACS, and HD-tRNS. Younger participants also followed a similar study design but without fixed-frequency beta-HD-tACS. Following 25 minutes of stimulation, participants were asked to complete a speech production task while their EEG activity was recorded. Both personalized tACS and tRNS significantly reduced reaction times compared to sham in younger and older adults. In older adults, personalized tACS and tRNS both effectively modulated speech motor network but exhibited distinct spatial profile. While personalized tACS produced more focal modulation, tRNS elicited broader modulation across distributed neural networks and engaged multiple oscillatory frequencies. Broadband and personalized stimulation approaches (personalized tACS and tRNS) were more effective than fixed-frequency beta-tACS in modulating behavioral and neural responses measured throughout a distributed speech-motor network in older adults.

neuroscience↗

bioRxiv2026

Early reactivation of medial temporal lobe neurons during emergence from propofol anesthesia in neurosurgical patients

Although much is known about the molecular and cellular effects of general anesthetics, the neural mechanisms underlying loss and recovery of consciousness during anesthesia remain elusive. We provide the first report of human single neuron activity recorded throughout emergence from general anesthesia. Following cessation of propofol infusion, emergence was assessed as motor response to verbal command (RC). In most cells, firing rates increased during the course of emergence. Analysis of pooled firing rate changes indicated a shift towards criticality during emergence and attractor states during and immediately following anesthesia and immediately after RC. Changes in activity in some regions occurred prior to overt RC, with shortest latencies in the hippocampus, parahippocampal gyrus, and amygdala, followed by cingulate and then insular cortex. This work reveals that brain activity in medial temporal regions may presage restoration of responsiveness in humans following general anesthesia.

neuroscience↗

bioRxiv2026

Transcranial magnetic stimulation to the dorsolateral prefrontal cortex modulates single-neuron activity in humans

Transcranial magnetic stimulation (TMS) to the dorsolateral prefrontal cortex (dlPFC) is an FDA-cleared treatment for depression, yet how cortical stimulation influences single neurons in deep brain circuits remains unknown. Using intracranial microelectrode recordings in four neurosurgical patients, we resolved single-neuron spikes as early as 8 ms from 185 single neurons after single-pulse left dlPFC TMS. TMS elicited time-locked firing responses in 46% of neurons across deep cortical and subcortical structures bilaterally. TMS facilitated putative interneuron spiking in striato-thalamic regions from [~]8 ms, peaking at [~]80-100 ms, and lasting to [~]1000 ms, while suppressing putative pyramidal cell spiking with a delayed and slower time course. Trial-by-trial single-neuron modulations were positively correlated with cortico-striato-thalamic network activity and anti-correlated with limbic network activity. These findings reveal that dlPFC TMS facilitates inhibitory firing in executive control networks while suppressing limbic excitatory drive, providing a cellular mechanism for how cortical stimulation modulates distributed brain networks.

neuroscience↗

bioRxiv2025

Distinct Temporal Patterns of Human Neural Firing in the Subthalamic Nucleus During Speech and Orofacial Movement

Clinical studies, along with electrophysiological findings, provide evidence that the subthalamic nucleus (STN) contributes to speech production. These studies have reported that the STN encodes diverse aspects of speech, comprising speech motor planning and execution, timing, and linguistic features such as phonetic content. However, none of these studies have included an orofacial non-speech motor task to evaluate speech-specificity of STN activity. Here, we examined the modulation of STN neurons while participants engaged in two speech tasks (sentence repetition and syllable repetition) as well as two non-speech orofacial movement tasks (jaw movement and tongue protrusion) in awake patients with Parkinsons disease undergoing deep brain stimulation implantation surgery. A total of 51 single- and multi-unit neural clusters were captured. A Poisson generalized linear model (GLM) was implemented to understand the temporal dynamics of STN activity. A larger proportion of clusters was modulated during speech (22%) than during orofacial movement (12%) and a substantial subset of STN neural clusters responded to overlapping speech and orofacial tasks (27%). The findings suggest that STN can encode both motor and linguistic aspects of speech production. Graphical abstractThe subthalamic nucleus (STN) shows neural modulation during speech production, a process which requires motor planning, execution, and phonological functions. By comparing STN spiking activity during speech tasks (sentence and syllable repetition) and non-speech orofacial tasks (jaw movement and tongue protrusion), we identified task-specific modulation patterns in STN neurons. The STN contains distinct neural populations engaged during speech and orofacial movements. Among all recorded clusters, 22% responded exclusively to speech, 12% exclusively to orofacial movements, 27% to both, and 39% were non-responsive. We demonstrated that STN activity at single- and multi-unit levels is specific to speech production and is influenced by task-specific motor and linguistic demands, highlighting a role for STN in integrating motor control and speech production. O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=108 SRC="FIGDIR/small/693637v1_ufig1.gif" ALT="Figure 1"> View larger version (27K): org.highwire.dtl.DTLVardef@1ec3c60org.highwire.dtl.DTLVardef@6fe46forg.highwire.dtl.DTLVardef@1fe733corg.highwire.dtl.DTLVardef@3f736b_HPS_FORMAT_FIGEXP M_FIG C_FIG

neuroscience↗

bioRxiv2025

Brain-wide single-neuron bases of working memory for sounds in humans

In order to understand the constantly changing acoustic world our brains must maintain elements of auditory scenes in memory. The neural mechanisms for this fundamental process remain unclear. Here, we report human intracranial recordings of 1269 single neurons while participants performed a non-verbal auditory working memory task that required adjustment of a tone frequency to match a target. We found neurons within hippocampus, insula and cingulate cortex, for which firing rates were modulated at different phases of the task, particularly during maintenance and adjustment. For the majority of the neurons modulated during maintenance there was a striking suppression of activity rather than increased activity. Across the entire neuronal population, state-space analysis demonstrated attractor-like states corresponding to different task phases. Behaviorally, more neurons were modulated at the beginning of the maintenance phase in trials when participants performed better. State-space analysis was consistent with greater attractor-like activity during the adjustment phase supporting better performance. These data support the existence of a widely distributed neural code for auditory working memory that determines related behavior based on attractor-like states.

neuroscience↗

bioRxiv2025

Reading specific memories from human neurons before and after sleep

The ability to retrieve a single episode encountered just once is a hallmark of human intelligence and episodic memory[1]. Yet, decoding a specific memory from neuronal activity in the human brain remains a formidable challenge. Here, we develop a transformer neural network model[2, 3] trained on neuronal spikes from intracranial microelectrodes recorded during a single viewing of an audiovisual episode. Combining spikes throughout the brain via cross-channel attention[4], capable of discovering neural patterns spread across brain regions and timescales, individual participant models predict vocalization of specific concepts such as persons or places during out-of-distribution memory recall. Brain regions differentially contribute to memory decoding before and after sleep. Models trained using only medial temporal lobe (MTL) spikes significantly decode concepts before but not after sleep, while models trained using only frontal cortex (FC) spikes decode concepts after but not before sleep, with better FC decoding relating to increased non-REM sleep. These findings suggest a system-wide distribution of information across neural populations that transforms over wake/sleep cycles[5]. Such decoding of internally generated memories suggests a path towards brain-computer interfaces to treat episodic memory disorders through enhancement or muting of specific memories.

neuroscience↗

bioRxiv2025

Brain Representations of Natural Sound Statistics

Natural sound textures (e.g., rain, crackling fire) are perceptually defined by time-averaged summary statistics. While previous studies have examined neural responses to natural sounds, little is known regarding the neural processing of the statistics underlying these sounds. To study neuronal correlates of these statistics, we measured brain responses to synthetic sound textures in which statistical structure was systematically varied while preserving the texture category. Using two fMRI experiments (males and females), we examined neural responses along the ascending auditory pathway, within auditory cortex and medial temporal lobe (MTL) regions implicated in pattern analysis. In Experiment 1, we parametrically varied the full set of texture statistics, creating sounds with different levels of naturalness. In Experiment 2, we selectively manipulated high-level statistics (cochlear skewness and kurtosis, cochlear and modulation correlations) while holding low-level statistics (cochlear mean and modulation power) constant. Increasing texture naturalness produced graded increases in BOLD responses across bilateral primary and nonprimary auditory cortex in both experiments, although overall responses were weaker in Experiment 2. This reduction suggests that low-level statistics contribute substantially to response magnitude, even though higher-order statistics are sufficient to elicit graded responses. We also observed modulation in MTL regions, including entorhinal cortex, in Experiment 1. Moreover, functional connectivity between hippocampus and auditory cortex increased for more degraded (less natural) textures, suggesting a modulatory rather than representational role for MTL in texture processing. Together, these findings show that sensitivity to texture statistics is distributed across the auditory cortex and highlight MTL-auditory interactions when texture structure is ambiguous. Significant StatementNatural sound textures such as rain or crackling fire are perceptually defined by time-averaged summary statistics that support efficient auditory perception, yet how the human brain represents these statistics remains unclear. Using fMRI and synthetic sound textures in which statistical structure was systematically manipulated, we found that both primary and nonprimary auditory cortex were sensitive to texture statistics, exhibiting a graded and distributed representation of these acoustic features. We also observed increased functional connectivity between the hippocampus and auditory regions when texture structure was degraded or the sounds were unnatural. Together, these results indicate that sound texture statistics are encoded across multiple levels of auditory cortex and further suggest a modulatory role for hippocampus under conditions of heightened perceptual uncertainty.

neuroscience↗

bioRxiv2025

Human insula neurons respond to simple sounds during passive listening

The insula is critical for integrating sensory information from the body with that arising from the environment. Although previous studies suggest that posterior insula is sensitive to sounds, auditory response properties of insula neurons have not previously been reported. Here, we provide the first report of a population of human single neuron data from the insula and provide comparative data from the primary auditory cortex, recorded intracranially from human participants during passive listening. In each condition, more than 330 single neurons were recorded in 11 participants. Almost a third of neurons in posterior insula and a smaller subset in anterior insula responded to simple tones and clicks. Responsive neurons were distributed throughout posterior and anterior insula and showed preferred frequency tuning. Onset latencies in the insula were similar to those in the primary auditory cortex but response durations were significantly shorter. Overall, these data highlight that insula neurons respond to auditory stimuli even in non-behaviorally relevant contexts and suggest an important contribution of audition to the postulated integrative functions of insular cortex.

neuroscience↗

bioRxiv2025

Brain bases for navigating acoustic features

Whether physical navigation shares neural substrates with mental travel in other behaviourally relevant domains is debated. With respect to sound, pure-tone working memory in humans elicits hippocampal as well as auditory cortical and inferior frontal activity, and rodent work suggests that hippocampal cells that usually track an animals physical location can also map to tone frequency when task-relevant. We generated a sound dimension based on the density of random-frequency tones in a stack, resulting in a percept ranging from low- ("beepy") to high-density ("noisy"). We established that unlike tone frequency, which listeners automatically associate with vertical position, this density dimension elicited no consistent spatial mapping. During functional magnetic resonance imaging, human participants held in mind the density of a series of tone stacks and, after a short maintenance period, adjusted further stacks to match the target ("navigation"). Density of the currently heard sound was represented most strongly in bilateral non-primary auditory cortex, specifically bilateral planum polare, while density of the maintained target was represented in right anterior hippocampus and left inferior temporal gyrus. Encoding and maintenance activity in bilateral hippocampus, inferior frontal gyrus, planum polare and posterior cingulate was positively associated with subsequent navigation success. Bilateral inferior frontal gyrus and hippocampus were among regions with elevated activity during adjustment, compared to a parity-judgment condition with closely matched acoustics and motor demands. Bilateral orbitofrontal cortex was more active when navigation was toward a target density than when participants adjusted density in a control condition with no particular target. We find that self-initiated travel along a non-spatial auditory dimension engages a brain system overlapping with that supporting physical navigation. Key PointsO_LIWork in rodents suggests that navigation in physical space and the active analysis of sounds share a neural substrate in the hippocampus, supporting the use of common computational mechanisms. C_LIO_LIWe examined the human brain system for navigation through an acoustic environment to a remembered target. C_LIO_LIIn addition to high-level auditory cortex we demonstrate involvement of the hippocampus in acoustic navigation along with other sites in frontal and cingulate cortex that also support physical navigation. C_LI

neuroscience↗

bioRxiv2024

Predicting speech-in-noise ability with static and dynamic auditory figure-ground analysis using structural equation modelling

Auditory figure-ground paradigms assess the ability to extract a foreground figure from a random background, a crucial part of central hearing. Previous studies have shown that the ability to extract static figures (with fixed frequencies) predicts real-life listening: speech-in-noise ability. In this study we assessed both fixed and dynamic figures: the latter comprised component frequencies that vary over time like natural speech. 159 participants (aged 18-79) with a range of peripheral hearing sensitivity were studied. We used hierarchal linear regression and structural equation modelling to examine how well speech-in-noise ability (for words and sentences) could be predicted by age, peripheral hearing, and static and dynamic figure-ground. Regression demonstrated that in addition to the audiogram and age, the low-frequency dynamic figure-ground accounted for significant variance of speech-in-noise, higher than the static figure-ground. The structural models showed that a combination of all types of figure-ground tasks predicted speech-in-noise with a higher effect size than the audiogram or age. Age influenced word perception in noise directly but sentence perception indirectly via effects on peripheral and central hearing. Overall, this study demonstrates that dynamic figure-ground explains more variance of real-life listening than static figure-ground, and the combination of both predicts real-life listening better than hearing sensitivity or age.

neuroscience↗

bioRxiv2024

British Version of the Iowa Test of Consonant Perception

The Iowa Test of Consonant Perception (ITCP) is a single-word closed-set speech- in-noise test with well-balanced phonetic features that provides a reliable testing option for real-world listening. Objectives. The current study aimed to establish a UK version of the test (B-ITCP) based on the British received pronunciation. Design. We conducted a validity test with 46 participants using the B-ITCP test, a sentence-in- noise test, and audiogram. Results. The B-ITCP demonstrated excellent test-retest reliability, cross-talker validity, and good convergent validity, consistent with the US results. Conclusions. These findings suggest that B-ITCP is a reliable measure of speech-in-noise perception, to facilitate comparative or combined studies in USA and UK. All materials (application and scripts) to run or construct the B-ITCP and ITCP are freely available online.

neuroscience↗

bioRxiv2022

Immediate neural network impact after the loss of a semantic hub

The human brain extracts meaning from the world using an extensive neural system for semantic knowledge. Whether such broadly distributed systems1-3 crucially depend on or can compensate for the loss of one of their highly interconnected hubs4-6 is controversial4. The strongest level of causal evidence for the role of a brain hub is to evaluate its acute network-level impact following disconnection and any rapid functional compensation that ensues. We report rare neurophysiological data from two patients who underwent awake intracranial recordings during a speech prediction task immediately before and after neurosurgical treatment that required disconnection of the left anterior temporal lobe (ATL), a crucial hub for semantic knowledge4-6. Informed by a predictive coding framework, we tested three sets of hypotheses including diaschisis causing disruption in interconnected sites7 and incomplete or complete compensation by other language-critical and speech processing sites8-10. Immediately after ATL disconnection, we observed highly specific neurophysiological alterations in the recorded fronto-temporal network, including abnormally magnified high gamma responses to the speech sounds in auditory cortex. We also observed evidence for rapid compensation, seen as focal increases in effective connectivity involving language-critical sites in the inferior frontal gyrus and speech processing sites in auditory cortex. However, compensation was incomplete, in part because after ATL disconnection speech prediction signals were depleted in auditory cortex. This study provides direct causal evidence for a semantic hub in the human brain and shows striking neural impact and a rapid attempt at compensation in a neural network after the loss of one of its hubs.

neuroscience↗

bioRxiv2021

Distribution of multi-unit pitch responses recorded intracranially from human auditory cortex

The perception of pitch requires the abstraction of stimulus properties related to the spectrotemporal structure of sound. Previous studies utilizing both animal electrophysiology and human imaging have indicated the presence of a center for pitch representation in the auditory cortex. Recent data from our own group - examining local field potentials (LFPs) in humans - indicate more widely distributed pitch-associated responses within the auditory cortex (Gander et al., 2019). To probe this with greater spatial resolution, we examined multi-unit activity related to three different auditory stimuli, in seven epilepsy patients who were implanted with high-impedance electrodes in auditory cortex for the clinical purpose of localizing seizures. The stimuli were regular-interval noise (RIN) with a pitch strength that is related to the temporal regularity, and pitch value determined by repetition rate, and harmonic complexes with missing fundamentals. We demonstrated increases in spiking activity in 69 of 104 (66%) responsive multiunit activity in auditory cortex due to pitch-associated stimuli. Importantly, these responses were distributed across the entire extent of Heschls gyrus (HG), in both primary and non-primary areas, rather than isolated to a specific region, and this finding was evident regardless of the stimulus presented. These findings are the first multi-unit pitch responses recorded from humans, and align with a recent study in macaques (Kikuchi et al., 2019) demonstrating that both local field potential and unit responses to pitch-inducing stimuli are distributed throughout auditory cortex. Significance StatementThe perception of pitch is a fundamental acoustic attribute that is mediated by the auditory system. Despite its importance, there is still debate as to the precise areas responsible for its encoding, which may be due to differences in the recording measures and choices of stimuli used in previous studies. Here, we present the first study to measure multi-unit pitch responses in the auditory cortices of intracranially-implanted humans. Importantly, we demonstrate reliable responses to three different pitch-inducing paradigms that are distributed throughout Heschls gyrus, rather than being localized to a particular region. These data provide a bridge across animal and human studies, and aid in our understanding of the processing of a critical attribute of acoustic stimuli.

neuroscience↗

bioRxiv2021

Neural correlates of individual differences in speech-in-noise performance in a large cohort of cochlear implant users

ObjectivesUnderstanding speech in noise (SiN) is a complex task that recruits multiple cortical subsystems. Individuals vary in their ability to understand SiN. This cannot be explained by simple peripheral hearing profiles, but recent work by our group (Kim et al., 2021, Neuroimage) highlighted central neural factors underlying the variance in SiN ability in normal hearing (NH) subjects. The current study examined neural predictors of speech-in-noise ability in a large cohort of cochlear-implant (CI) users, with the long-term goal of developing a simple electrophysiological correlate that could be implemented in clinics. DesignWe recorded electroencephalography (EEG) in 114 post-lingually deafened CI users while they completed the California Consonant Test (CCT): a word-in-noise task. In many subjects, data were also collected on two other commonly used clinical measures of speech perception: a word-in-quiet task (Consonant-Nucleus-Consonant [CNC]) word and a sentence-in-noise task (AzBio sentences). Neural activity was assessed at a single vertex electrode (Cz), to maximize generalizability to clinical situations. The N1-P2 complex of event-related potentials (ERPs) at this location were included in multiple linear regression analyses, along with several other demographic and hearing factors as predictors of speech in noise performance. ResultsIn general, there was a good agreement between the scores on the three speech perception tasks. ERP amplitudes did not predict AzBio performance which was predicted by the duration of device use, low-frequency hearing thresholds, and age. However, ERP amplitudes were strong predictors for performance for both word recognition tasks: the CCT (which was conducted simultaneously with EEG recording), and the CNC (conducted offline). These correlations held even after accounting for known predictors of performance including residual low-frequency hearing thresholds. In CI-users, better performance was predicted by an increased cortical response to the target word, in contrast to previous reports in normal-hearing subjects in whom speech perception ability was accounted for by the ability to suppress noise. ConclusionsThese data indicate a neurophysiological correlate of speech-in-noise performance that can be relatively easily captured within the clinic, thereby revealing a richer profile of an individuals hearing performance than shown by psychoacoustic measures alone. These results also highlight important differences between sentence and word recognition measures of performance and suggest that individual differences in these measures may be underwritten by different mechanisms. Finally, the contrast with prior reports of NH listeners in the same task suggests CI-users performance may be explained by a different weighting of neural processes than NH listeners.

neuroscience↗