Biology subjects

Baciu, M.

Publications and source records attributed to Baciu, M..

11 recordsLinked to original sources

bioRxiv2026

Exploring Healthy Neurocognitive Ageing with Deep Learning Interpretability Methods

Healthy ageing reorganises large-scale functional brain networks. Using resting-state fMRI from 615 adults aged 18 to 88 years in Cam-CAN. We combined raw functional connectivity analyses, graph-theoretical topology, Ridge and MLP age prediction and their attributions in order to understand how functional brain organisation varies with age and relates to cognitive performance. Both models predicted age accurately, with a modest advantage for the MLP. Within-network FC decreased with age, while between-network FC and participation increased. Topological analyses showed increased participation and a redistribution of nodes across connector, provincial, satellite, and peripheral roles. The MLP SmoothGrad attributions emphasised high-participation and weakly specialised nodes, whereas counterfactual tests showed that replacing CON-SMN, CON-DMN and DMN-FPN connections with a young-adult template produced the largest reductions in predicted age. Brain-cognition PLS linked these mechanisms to a broad cognitive-performance axis and a secondary semantic/connectomic axis. Together, these findings suggest that healthy ageing involves a redistribution of functional organisation that extends beyond a single DMN-FPN axis towards a broader DMN-CON-SMN-FPN configuration.

neuroscience↗

bioRxiv2026

MEG State Dynamics of Sentence Generation: Evidence for a Compensatory Chunking Mechanism in Healthy Aging

Healthy aging is accompanied by subtle difficulties in language production. While behavioral and neuroimaging studies suggest that older adults rely on acute semantic access to maintain language abilities, the underlying neurophysiological mechanisms remain poorly understood. In particular, it is still unclear how large-scale brain dynamics reorganize to support naturalistic sentence generation with age. In this study, we investigated the spatiotemporal brain-state dynamics during covert sentence generation (GE2REC protocol) in younger and older adults using magnetoencephalography (MEG). Source-reconstructed MEG signals were analyzed using a Hidden Markov Model which identified five recurrent brain states, encompassing language-semantic, language-control, sensorimotor, and visual domains. Latent modeling was then used to relate the spectral and temporal properties of these brain states to age and language performance. Spectrally, older adults appear to redistribute oscillatory activity from sensorimotor-related states toward semantic-related states across alpha, beta, and low-gamma frequency bands. Temporally, older adults exhibit a more segmented processing sequence between semantic and sensorimotor processing which interfaces with visuo-posterior processing. These changes robustly covaried with age and better verbal fluency (semantic & lexical). Taken together, these results suggest that the older adult brain undergoes a coordinated time-frequency reorganization to support sentence production. Individuals likely establish an embodied semantic strategy in aging that involves "chunking" the processing stages of sentence production via visuo-posterior information processing. We speculate that this may help shape a resource-efficient, predictive route for complex cognition in older adulthood.

neuroscience↗

bioRxiv2025

Structure-Function Dynamics in Healthy Cognitive Aging: A Graph Signal Processing Approach

Network neuroscience has significantly advanced our understanding of how structural and functional connectivity evolve during healthy neurocognitive aging. Yet, integrative studies linking structural and functional brain organization with cognitive performance remain relatively limited. In this study, we analyzed resting-state functional MRI and diffusion-weighted imaging data from 600 healthy adults aged 18 to 88, drawn from the CamCAN dataset. Using a graph signal processing framework, we investigated how structural connectivity constrains functional brain signals across the adult lifespan. Our results reveal that control and semantic cognitive systems exhibit distinct age-related patterns of structure-function reorganization, potentially reflecting a shift in integrative processing during midlife. Notably, our findings suggest that structurally-coupled sensorimotor integration plays a crucial role for regulating these systems. Until midlife, it accompanies structurally-decoupled activity in transmodal cortices to sustain cognitive control. In parallel, it likely supports the formation of embodied internal models that leverage more structurally-decoupled semantic processes, thus contributing to maintain lexical production skills for longer in older adulthood. Taken together, our study offer new multimodal insights into how sensory-driven processes help reconfigure the healthy aging brain to support controlled semantic cognition.

neuroscience↗

bioRxiv2025

A Neurocognitive Shift in Midlife: Linking Cognitive Flexibility and Functional-Metabolic Adaptation with the SENECA model

Understanding how the older adult brain sustains cognitive flexibility remains a central question in aging research. Here, we analyzed resting-state fMRI data from the population-based CamCAN database (N = 628; age 18-88) and applied structural balance theory to measure functional network energy, a graph-theoretical proxy of network flexibility. In line with the SENECA model, our findings highlight midlife as a critical transition period: network energy is redistributed along the sensory-transmodal hierarchy, shifting from higher-level networks (DMN-FPN) to lower-level networks (SMN, CON, Auditory, Visual, Language). This reorganization (i) helps preserve a global wiring economy across the lifespan, hinting at an allostatic mechanism (i.e., stability through change); and (ii) may support embodied semantic strategies in older adulthood, leveraging more predictive processing to sustain cognitive flexibility at lower costs. Taken together, our study reframes healthy neurocognitive aging as an allostatic process and provides a reference for extending the SENECA model to metabolism and neuropathology.

neuroscience↗

bioRxiv2024

Lifespan Oscillatory Dynamics in Lexical Production: A Population-based MEG Resting-State Analysis

Lexical production remains relatively preserved across the lifespan, but cognitive control demands increase with age to support efficient semantic access. It suggests a domain-general and a language-specific component. Current neurocognitive models suggest the Default Mode Network (DMN) may drive the interplay between these components, impacting the trajectory of production performance with a pivotal shift around midlife. However, the corresponding time-varying architecture still needs clarification. Here, we leveraged MEG resting-state data from healthy adults aged 18-88 from a CamCAN population-based sample. We found that DMN temporal dynamics shift from anterior-ventral to posterior-dorsal states until midlife to mitigate word-finding challenges. Similarly, sensorimotor integration along this posterior path enhances cross-talk with lower-level circuitry as the dynamic information flow with more anterior, higher-order cognitive states gets compromised. It suggests a bottom-up, exploitation-based form of cognitive control in the aging brain, highlighting the interplay between abstraction, control, and perceptive-motor systems in preserving lexical production. HighlightsO_LIMidlife is a pivotal period for time-varying functional connectivity C_LIO_LIDMN activation and deactivation drive the resting-state oscillatory architecture C_LIO_LIEnhanced posterior DMN temporal dynamics mitigates lexical production decline C_LI

neuroscience↗

bioRxiv2024

Dynamics of White Matter Architecture in Lexical Production among Middle-Aged Adults

This study aimed to elucidate the white matter changes associated with lexical production (LP) difficulties that typically emerge in middle age, resulting in increased naming latencies. To delay the onset of LP decline, middle-aged adults may rely on domain-general (DG) and language-specific (LS) compensatory mechanisms as proposed by the LARA model (Lexical Access and Retrieval in Aging). However, our knowledge of the white matter changes supporting these mechanisms remains incomplete. Based on a sample of 155 middle-aged adults from the CAMCAN cohort, we combined dimensionality reduction techniques with multivariate statistical methods to jointly examine the relationships between diffusion-weighted imaging and LP-related neuropsychological data. Our findings (i) show that midlife constitutes a pivotal period marked by a discontinuity in brain structure within distributed networks within dorsal, ventral, and anterior cortico-subcortical pathways, and (ii) reveal that this discontinuity signals a neurocognitive transition around age 53-54, marking the onset of LP decline. Indeed, our results propose that middle-aged adults may initially adopt a "semantic strategy" to compensate for initial LP challenges. Still, this strategy may be compromised when late middle-aged adults (age 55-60) lose the ability to exert cognitive control over semantic representations (i.e., reduced semantic control). In summary, our study advances our comprehension of brain structure changes that underpin the neurocognitive profile of LP in middle age. Specifically, we underscore the importance of considering the interplay between DG and LS processes when studying the trajectory of LP performance in healthy aging. Furthermore, these findings offer valuable insights into identifying predictive biomarkers related to the compensatory dynamics observed in midlife, which can help understand language-related neurodegenerative pathologies. HighlightsO_LIMidlife constitutes a pivotal period characterized by a discontinuity in brain structure. C_LIO_LIEarly middle-aged adults (age 45-55) adopt a "semantic strategy" to facilitate semantic access and sustain lexical production (LP) performances. C_LIO_LILate middle-aged adults (age 55-60) gradually lose the ability to exert cognitive control over semantic representations, marking the onset of LP decline. C_LI O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=105 SRC="FIGDIR/small/579645v1_ufig1.gif" ALT="Figure 1"> View larger version (28K): org.highwire.dtl.DTLVardef@85346dorg.highwire.dtl.DTLVardef@13515b7org.highwire.dtl.DTLVardef@24a1fforg.highwire.dtl.DTLVardef@1078f4c_HPS_FORMAT_FIGEXP M_FIG C_FIG

neuroscience↗

bioRxiv2023

When Age Tips the Balance: a Dual Mechanism Affecting Hemispheric Specialization for Language

Aging leads to neuroadaptations, often reducing specificity in functional brain responses. However, the extent to which functional specialization of brain hemispheres in the Language-and-Memory Network regions changes with adulthood remains unclear. In a cohort of healthy adults, we provide evidence that aging is linked to shifts in the cortical organizations lateralization, unveiling two age-related asymmetry patterns. The first pattern indicates a lateralization shift in language regions, with aging rendering language functions more bilateral. The second pattern reveals increased specialization in memory regions within the left hemisphere as we age. Manifesting around midlife, these shifts correlate with declines in language production performance. Our findings offer new insights into how functional brain asymmetries impact language proficiency and underscore brain plasticity in aging, providing a dynamic view of the aging brains functional architecture.

neuroscience↗

bioRxiv2023

Modeling the Neurocognitive Dynamics of Language across the Lifespan

Healthy aging is associated with a heterogeneous decline across cognitive functions, typically observed between language comprehension and language production (LP). Examining resting-state fMRI and neuropsychological data from 628 healthy adults (age 18-88) from the CamCAN cohort, we performed state-of-the-art graph theoretical analysis to uncover the neural mechanisms underlying this variability. At the cognitive level, our findings suggest that LP is not an isolated function but is modulated throughout the lifespan by the extent of inter-cognitive synergy between semantic and domain-general processes. At the cerebral level, we show that DMN (Default Mode Network) suppression coupled with FPN (Fronto-Parietal Network) integration is the way for the brain to compensate for the effects of dedifferentiation at a minimal cost, efficiently mitigating the age-related decline in LP. Relatedly, reduced DMN suppression in midlife could compromise the ability to manage the cost of FPN integration. This may prompt older adults to adopt a more cost-efficient compensatory strategy that maintains global homeostasis at the expense of LP performances. Taken together, we propose that midlife represents a critical neurocognitive juncture that signifies the onset of LP decline, as older adults gradually lose control over semantic representations. We summarize our findings in a novel SENECA model (Synergistic, Economical, Nonlinear, Emergent, Cognitive Aging), integrating connectomic and cognitive dimensions within a complex system perspective. HighlightsO_LILexical production (LP) relies on the interplay between domain-general and semantic processes throughout life. C_LIO_LIDMN (Default Mode Network) suppression cooperates with FPN (Fronto-Parietal Network) integration to maintain LP performance at a minimal cost. C_LIO_LIMidlife marks a neurocognitive shift, with reduced DMN suppression prompting a more cost-efficient compensatory strategy that prioritizes homeostasis over LP performance. C_LI O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=135 SRC="FIGDIR/small/547510v2_ufig1.gif" ALT="Figure 1"> View larger version (36K): org.highwire.dtl.DTLVardef@85939aorg.highwire.dtl.DTLVardef@19d8bbcorg.highwire.dtl.DTLVardef@40f5e8org.highwire.dtl.DTLVardef@d0713f_HPS_FORMAT_FIGEXP M_FIG C_FIG

neuroscience↗

bioRxiv2022

Unraveling the functional attributes of the language connectome: crucial subnetworks, flexibility and variability.

Language processing is a highly integrative function, intertwining linguistic operations (processing the language code intentionally used for communication) and extra-linguistic processes (e.g., attention monitoring, predictive inference, long-term memory). This synergetic cognitive architecture requires a distributed and specialized neural substrate. Brain systems have mostly been examined at rest. However, task-related functional connectivity provides additional and valuable information about how information is processed when various cognitive states are involved. We gathered thirteen language fMRI tasks in a unique database of one hundred and fifty neurotypical adults (InLang database). The tasks were designed to assess a wide range of linguistic processes and subprocesses. From this database, we applied network theory as a computational tool to model the task-related functional connectome of language (LANG). The organization of this data-driven neurocognitive atlas of language is examined at multiple levels, uncovering its major components (or crucial subnetworks) and its anatomical and functional correlates. Furthermore, we estimate its reconfiguration as a function of linguistic demand (flexibility), or several factors such as age or gender (variability). By accounting for the multifaceted nature of language and modulating factors, this study can contribute to enrich and refine existing neurocognitive models of language. The LANG atlas can also be considered as a reference for comparative or clinical studies, involving a variety of patients and conditions.

neuroscience↗

bioRxiv2019

Interactive mapping of language and memory with the GE2REC protocol

Previous studies have highlighted the importance of considering cognitive functions in a dynamic and interactive perspective and multiple evidence was brought for a language and memory interaction. In this study performed in healthy participants, we developed a new protocol entitled GE2REC that interactively accesses the neural representation of language-and-memory network. This protocol consists of three runs related to each other, providing a link between tasks, in order to assure an interactive measure of linguistic and episodic memory processes. GE2REC consists of a sentence generation (GE) in auditory modality and two recollecting (2REC) memory tasks, one recognition performed in visual modality, and another one recall performed in auditory modality. Its efficiency was evaluated in 20 healthy volunteers using a 3T MR imager. Our results corroborate the ability of GE2REC to robustly activate a fronto-temporo-parietal language network as well as temporal mesial, prefrontal and parietal cortices during encoding and recognition. GE2REC is useful because: (a) requires simultaneous and interactive language-and-memory processes and jointly maps their neural basis; (b) explores encoding and retrieval, managing to elicit activation of mesial temporal structures; (c) is short and easy to perform, hence being suitable for more restrictive settings, and (d) has an ecological dimension of tasks and stimuli. Overall, GE2REC can provide valuable information in terms of the practical foundation of exploration language and memory interconnection.

neuroscience↗

bioRxiv2017

Frequency-Resolved Correlates Of Visual Object Recognition In Human Brain Revealed By Deep Convolutional Neural Networks

Previous work demonstrated a direct correspondence between the hierarchy of the human visual areas and layers of deep convolutional neural networks (DCNN) trained on visual object recognition. We used DCNNs to investigate which frequency bands correlate with feature transformations of increasing complexity along the ventral visual pathway. By capitalizing on intracranial depth recordings from 100 patients and 11293 electrodes we assessed the alignment between the DCNN and signals at different frequency bands in different time windows. We found that gamma activity, especially in the low gamma-band (30 - 70 Hz), matched the increasing complexity of visual feature representations in the DCNN. These findings show that the activity of the DCNN captures the essential characteristics of biological object recognition not only in space and time, but also in the frequency domain. These results also demonstrate the potential that modern artificial intelligence algorithms have in advancing our understanding of the brain.\n\nSignificance StatementRecent advances in the field of artificial intelligence have revealed principles about neural processing, in particular about vision. Previous works have demonstrated a direct correspondence between the hierarchy of human visual areas and layers of deep convolutional neural networks (DCNNs), suggesting that DCNN is a good model of visual object recognition in primate brain. Studying intracranial depth recordings allowed us to extend previous works by assessing when and at which frequency bands the activity of the visual system corresponds to the DCNN. Our key finding is that signals in gamma frequencies along the ventral visual pathway are aligned with the layers of DCNN. Gamma frequencies play a major role in transforming visual input to coherent object representations.

neuroscience↗