Cognitive Neurodynamics

Neuromarketing mainly seeks to enhance the prediction of marketing stimuli success, such as commercials and movie trailers, by integrating neurophysiological measures with traditional behavioral measures. In the current study, the authors tested whether neural activity could predict consumer valuation and how preferences change after watching commercials, both immediately and over time. Participants (n=161) watched images of products followed by commercials advertising those products while their neural activity was recorded using electroencephalograph (EEG). They watched the product images again a week later without EEG recordings. Immediately after each stimulus exposure, participants stated their willingness to pay (WTP) for the product and how much they liked the commercial. Standard EEG measures showed weak and inconsistent relationships with behavior and yielded near-chance predictions. In contrast, deep learning models applied to the raw neural data achieved substantially higher predictive accuracy, successfully predicting both immediate and delayed WTP and ad liking. Importantly, the authors were able to successfully predict preferences for new participants and/or new products the models were not trained on. However, accuracy declined as generalization demands increased. Moreover, prediction performance declined as value differences narrow. Together, these findings show that neural responses carry reliable and temporally persistent information about consumer valuation and its evolution over time.

Recurrent cortical circuits are hypothesized to perform Bayesian posterior sampling through their stochastic dynamics. Yet identifying which sampling algorithm a circuit intrinsically implements, as opposed to mapping manually designing algorithms onto circuits, requires rigorous theoretical analysis of the biological circuit dynamics, and remains a fundamental open challenge because of the analytical intractability of nonlinear recurrent dynamics. Here, we tackle this challenge through comprehensive theoretical analysis of analytically tractable nonlinear recurrent circuit dynamics used in neuroscience research-- the continuous attractor network (CAN). Surprisingly, we find that the circuit, without any non-Gaussian components in its connectivity or internal variability, intrinsically emerges heavy-tailed subspaces dynamics described as scale-mixture Langevin posterior sampling. The mechanism emerges from the circuits own nonlinear dynamics: divisive normalization renders the overall population activation Gamma-distributed which multiplicatively modulates the time constant of the Langevin sampling in the stimulus feature subspace. Eventually, this scale-mixture structure produces a heavy-tailed (Student-t) sampling step size distribution, significantly accelerating the sampling speed. Moreover, the scale-mixture Langevin sampling approaches a Levy-like regime in the weak input limit. Our results analytically identify an intrinsic circuit algorithm embedded in subspaces of nonlinear recurrent dynamics, and establish a minimal circuit mechanism for endogenous heavy-tail statistics without any explicit non-Gaussian components.

The discrepancy between chronological age and predicted brain age, derived from neuroimaging data, serves as a biomarker for neurological health. However, the exact electroencephalography (EEG) mechanisms contributing to the variation of brain age are yet to be understood. This paper comprehensively investigated the relationship between brain age estimated from EEG-based functional connectivity and cognitive or mental manifestations of healthy adults. A mean absolute error (MAE) of 5.97 and R-squared (R2) of 0.85 were reported for training data, while an MAE of 10.73 and R2 of 0.56 were reported for the prediction data. Increased brain age was found to be negatively related to the scores of memory and attention tasks. Most young adults (19 of 26) suffering from mental symptoms such as addictions, depression, phobia, and anorexia nervosa demonstrated higher brain age than their chronological age. Alcohol consumption was also significantly correlated to the difference between brain and chronological ages. The functional connectivity of frontal and central motor regions was found to be most distinguishing between individuals with high and low brain ages. These brain regions were significantly correlated with brain age but not chronological age, suggesting a relatively stronger relationship with biological age. Ablation studies further confirmed the importance of the frontal brain and all bands EEG in the prediction of brain age. This work advances the understanding of EEG mechanisms contributing to brain age and could provide an interpretation for physiological or psychological conditions associated with brain age.

Alterations in impulsivity characterize several neuropsychiatric disorders, however the neural processes that underlie an impulsive choice have not been identified. A recent body of literature suggests that there may be generalizable aspects of neural representations across animals. Whether this is also true for dynamic rather than just static (like geometrical) properties of neural representations, and specifically in the context of impulsive behaviors, is less clear. Therefore, the goal of this study is to characterize generalizable features of latent dynamics that disambiguate an impulsive from non-impulsive choice across many different data sets and animals. A deep learning-based latent factor model was constructed from neural recordings that were acquired in rat anterior cingulate cortex while performing a delay discounting task, which is commonly used to assess impulsivity. Prior work shows this brain region is critical for the performance of this task. We leveraged piecewise linear recurrent neural networks (PLRNNs) to perform a nonlinear latent factor analysis across animals and experimental sessions. Our PLRNN explained a substantially larger percent of variance in fewer dimensions than linear approaches. We cross-validated performance of the model on test data and following the disruption of recurrent connectivity or auto-regressive latent dynamics. These tests confirmed that our fitting procedure consistently embedded generalizable neuronal dynamics within the PLRNN. Further, we found that the latent dynamics exhibited strong rotational features and progressed faster for impulsive choices. In summary, our model captures properties of the latent dynamics that are consistent across multiple data sets and disambiguate impulsive choices.

Musical aesthetic experience can reveal students' cognitive, aesthetic, and emotional responses to music. However, how professional vocal training alters brain functions in female college students with varying levels of musicianship to promote these is unknown. Here we applied fMRI, psychological measures and Goldsmiths musical sophistication index (Gold-MSI) assessments to study 27 non-music major female college students consisting of instrument group (normal pitch accuracy with instrumental training), general group (normal pitch accuracy without instrumental training) and poor-pitch group (poor pitch accuracy and without instrumental training). After 1 year of one-on-one instructional practice, we found that professional vocal training enhanced information processing related to musical cognition and pleasure during musical aesthetic experiences across all three groups, involving improvements in motor control of vocalization, auditory feedback processing, musical memory and attention together with musical pleasure. Specifically, instrument group showed activation of reward circuits, general group exhibited activation of laryngeal motor control circuits, and poor-pitch group displayed activation of sensory/sensorimotor/motor control circuits. Both pre- and post-training, the instrument group exhibited higher connectivity than the other two groups in the left amygdala-dorsal anterior cingulate cortex (dACC) and the right posterior superior temporal gyrus (pSTG)-left insula pathway. After professional vocal training, the general group showed higher connectivity in the left amygdala-dACC, whereas the poor-pitch group demonstrated increased connectivity in the right pSTG-left insula and dACC-nucleus accumbens pathways. Surprisingly, professional vocal training activated the bed nucleus of the stria terminalis (BNST), a subcortical emotion center, and connectivity with the medial prefrontal cortex associated with musical information processing. Thus, vocal training can significantly improve musical cognition and aesthetic pleasure, and facilitates neural modulation in the brain's emotional centers.

Brain network organization emerges from ongoing network dynamics that continuously create, maintain, and reorganize connectivity patterns. At both the structural and functional levels, these dynamics shape network evolution throughout the lifespan. Here, we introduce a data-driven framework based on Stochastic Actor-Oriented Models (SAOMs), which model longitudinal network changes as a process of local edge formation and dissolution driven by local-connectivity factors. Applied to functional and structural brain networks from an aging cohort, the resulting models recovered a set of interacting mechanisms that collectively promote hub formation, short-range connectivity, and modular organization. These factors were derived through a novel data-driven factor-selection procedure rather than specified a priori, allowing network mechanisms underlying longitudinal change to be identified directly from observations. Functional networks exhibited substantially higher rates of change than structural networks, suggesting more frequent opportunities for network maintenance and reorganization. Structural networks additionally showed despecification in the hub-forming mechanisms with cognitive impairment. These results support a view of brain network organization as an actively maintained process emerging from ongoing local connectivity dynamics. More broadly, this framework provides a general approach for studying longitudinal network dynamics through the estimation of network-generating mechanisms over time.

The -Bonacci weighted sequence space {ell}(,) is introduced for [&ge;] 2, 1 [&le;] < {infty}, and > 0). It is proved to be a Banach space which is isometrically isomorphic to {ell}, with an unconditional Schauder basis. The space is also reflexive and uniformly convex. Its strict inclusion chain, continuous dual, and asymptotic truncation estimate of order - / are also established. Based on this framework, Scaled &#119898-BonacciNet is proposed as a lightweight CNN that rescales Fibonacci-Net channels using the dominant &#119898-Bonacci root and employs depth wise separable convolutions. Tribonacci-Net achieved 96.09% accuracy and a 0.9581 macro F1-score on the 15-class Brain MRI dataset using 0.85 million parameters and 0.366 GFLOPs. Pentanacci-Net attained 95.87% accuracy on the 44-class dataset while reducing parameters and FLOPs by 47.10% and 44.48%, respectively. On the CKD ultrasound dataset, Hexanacci-Net achieved 83.75% accuracy and a 0.7941 macro F1-score, with corresponding reductions of 50.37% and 45.10%. These findings support dominant-root scaling as a structured approach to lightweight CNN design.

Motor effort has been shown to bias perceptual decisions toward less effortful choices, but it remains unknown whether this bias already takes effect during the early stages of decision formation through preparatory attentional allocation or sensory processing, or during later-stage evidence accumulation preceding response execution. This study investigated when motor effort becomes integrated into decision formation by focusing on the temporal evolution of neural processing during perceptual decision-making. While 64-channel EEG was recorded, twenty-four healthy adults performed an orientation discrimination task in which two Gabor patches appeared simultaneously in the left and right visual hemifield and decisions were reported using two pressure-sensitive response buttons requiring different levels of force. We found that motor effort modulated EEG activity at both the preparatory and later decision stages: pre-stimulus alpha power was reduced over the occipito-parietal cortex contralateral to the low-effort option, and mean centro-parietal positivity (CPP) amplitude preceding response execution was greater over centro-parietal cortex for low-effort choices, while leaving early sensory representations intact as indicated by no differences in P1 or N1 amplitude following stimulus onset. Trial-by-trial CPP amplitude preceding response execution increased with the likelihood of selecting the low-effort option, whereas the pre-stimulus alpha modulation was not associated with behavioral decision bias. These findings suggest that motor effort biases perceptual decisions primarily during later-stage evidence accumulation, thereby providing insight into the neural basis of embodied perceptual decision making.

We investigated how auditory and visual presentation of small numerical magnitudes and non- numerical magnitudes are passively processed at the neural level. Earlier research suggests automatic processing of quantities occurs rapidly in the brain. However, there are no studies on passive, pre-attentive change detection of magnitudes utilizing mismatch negativity (MMN) across sensory modalities in same participants, important to understand amodal processing of magnitudes. We recorded EEG in passive auditory and visual multi-feature oddball and control paradigms from the same adult individuals (N=104/105, respectively). In the auditory paradigm, the standard stimulus was a 2-part sound, while the deviant stimuli varied in frequency, intensity, gap length, and number of sounds (1-part and 3-part stimuli). In the visual paradigm, the standard stimulus was a dot array containing two medium-sized dots, while the deviant stimuli varied in numerosity (1 or 3 dots) and dot size (small or large). The results for the auditory paradigm showed significant MMN across all features, including comparisons with deviant-alone control stimuli, except for the 2-part vs. 1-part contrast. The results for the visual paradigm suggested differences in processing across all features apart from large and medium-size stimuli. These results replicate earlier MMN findings that the brain distinguishes pre-attentively changes in sound frequency, amplitude and gap. Auditorily presented small numerical magnitudes also showed pre-attentive processing but less robustly to the contrast 2 vs. 1. Visual pre-attentive processing of small numerical magnitudes was also observed. Together these findings could be interpreted to support the existence of abstract, pre- attentive processing of numerical magnitudes. Highlights* Auditory and visual processing of numerical and non-numerical magnitudes was examined with EEG * Auditory non-numerical magnitudes and medium vs. small dots replicated earlier MMN studies * aMMN and vMMN showed pre-attentive processing of changes in small numerosities 2 vs. 3 and 2 vs. 1 * 2 vs. 1 tone processing was less robust than 2 vs. 3 or the processing of the visual contrasts * The brain processes changes in small non-symbolic magnitudes pre-attentively

Working memory (WM) and decision-making (DM) are closely linked cognitive processes that rely on frontoparietal regions, yet their distinct and shared neural dynamics remain unclear. Using intracranial EEG recordings from 24 human subjects performing a delayed match-to-sample task, we examined high-frequency broadband (HFB) activity to dissociate WM- and DM-related processes with high spatiotemporal resolution. Principal component analysis of frontoparietal HFB activity revealed multiple temporally distinct patterns spanning stimulus presentation and delay phases. The most dominant component reflected sustained WM-related activity across task phases, consistent with persistent maintenance of internal representations. Additional components captured transient WM encoding across trial phases, as well as both stimulus-linked and ramping DM-related activity suggesting dual early and late stages of evaluative processing. These processes were distributed across frontoparietal regions but showed systematic differences in anatomical focus, with WM-related activity concentrated in dorsolateral prefrontal cortex (PFC) and DM-related activity concentrated in ventrolateral PFC, while inferior parietal cortex preferentially tracked the behavioral significance of incoming stimuli and superior parietal cortex tracked the appearance of new information that may require updating ongoing processing. Together, these findings demonstrate that WM and DM processes are separable yet interleaved over time within distributed cortical networks, revealing coordinated but functionally distinct contributions to goal-directed behavior.

Background: Neuroanatomy poses a significant challenge for Psychology students due to its spatial and conceptual complexity. Educational approaches that enhance the relevance and visualization of neuroanatomical content may improve students learning experiences. This study implemented a tractography-based activity focused on the virtual dissection of the arcuate fasciculus, a major white matter pathway, in undergraduate Psychology students and examined the relationships between perceived learning and students perceptions of utility, difficulty and handling, and organizational aspects of the activity. Methods: First-year undergraduate Psychology students participated in a two-session tractography-based activity combining instruction on white matter anatomy and diffusion tractography with a hands-on virtual dissection of the arcuate fasciculus using research-grade software routinely employed in neuroscience research. Following the activity, students completed an anonymous questionnaire assessing perceived learning, utility, difficulty and handling, and organizational aspects of the activity. Pearson correlations, multiple regression analyses, and relative importance analyses were performed. Results: Sixty-eight students completed the questionnaire. Students reported generally positive perceptions of the activity across the evaluated dimensions, with perceived learning receiving the highest mean score (M = 3.44, SD = .85). Perceived utility showed the strongest association with perceived learning (r = .64, p < .001). The regression model explained 41% of the variance in perceived learning (R2 = .41, adjusted R2 = .38, p < .001). Perceived utility was the only significant predictor in the model ({beta} = .59, p = .001), accounting for 67.9% of the explained variance. Conclusions: The findings support the feasibility of integrating authentic neuroimaging tools into undergraduate neuroanatomy teaching. Students who perceived the activity as more useful also reported higher perceived learning outcomes, with perceived utility emerging as the strongest predictor of perceived learning. In contrast, perceived difficulty and handling, and organizational aspects did not make significant independent contributions. These results suggest that students perceptions of educational relevance may play an important role in technology-enhanced STEM learning experiences.

Face recognition ability varies enormously across humans. Individuals with face-blindness (prosopagnosia) struggle to recognize even close family members while super-recognizers can identify strangers with exceptional accuracy. Where do ANNs fall within the human face recognition spectrum? Here, we administered the standardized tests used to characterize human face recognition ability to a diverse set of models, allowing us to contextualize a model's performance within the distribution of human behavior. We found that the majority 55% of models to be classified as face-blind and that even the best face-trained models do not cross the human threshold to be considered a super-recognizer. Interrogating the internal representations of these models showed that models that performed well on standardized face recognition tasks were more identity-selective and viewpoint-invariant. We also found that low-performing models did contain some identity information in independent representational subspaces. Removing viewpoint-dependent subspace improved face recognition abilities in 49 of the 53 models tested. Targeted unit ablations further identified opposing contributions, with viewpoint-dependent units disrupting identity coding and viewpoint-tolerant units supporting it. Together, our results show that most AI models are face-blind with worse face recognition ability than humans, and demonstrate how differences across AI models can be used generate testable hypotheses about the computational basis of human face recognition in humans which can then be probed in future studies.

Nonlinear neural networks develop structured internal representations, yet how their geometry is determined by the tasks being learned remains poorly understood. Here, we develop an analytical framework for piecewise-linear feedforward and recurrent networks that links learning, activation-region structure, and neural selectivity. We show that, at any stationary point of gradient-aligned learning, a nonlinear network decomposes into local linear regressions over its activation regions. Deviations from the corresponding least-squares solutions are jointly constrained by network weights shared across regions and vanish in the low-error regime, yielding an approximate piecewise least-squares decomposition of the task. This structure admits a decision-tree interpretation, which we recover numerically by fitting trees to predict network activation patterns from the input. We further derive how the regions in which neurons are active determine the task statistics they capture, thereby organizing neural selectivity into distinct subpopulations. The resulting predictions closely match the selectivity geometry observed in simulated networks and two empirical neural datasets. Finally, we show that neural baseline regulates activation pattern diversity, placing networks along a continuum between coarse, generalizing representations and fine-grained, expressive representations. Together, these results establish activation regions as a unifying framework for describing how nonlinear networks decompose task structure into local computations and how those computations can be revealed through neural activity.

Astrocytes actively regulate synaptic transmission and neuronal excitability, yet their role in orchestrating macroscopic cortical network regimes and slow-wave oscillations remains an active area of reasearch. This study investigates how bidirectional neuron astrocyte interactions shape emergent population dynamics using a computational network model of excitatory and inhibitory neurons coupled to an astrocyte. The results identify astrocytic feedback topology, rather than astrocytic coupling strength alone, as a key determinant of emergent cortical network dynamics. By systematically dissecting pathway-specific connectivity, it has been shown that the neuronal population driving astrocytic activation and the neuronal population receiving gliotransmission jointly determine whether the network occupies asynchronous irregular (AI), synchronous irregular (SI), synchronous regular(SR), asynchronous regular(AR) or quiescent regimes.Directing gliotransmission selectively onto excitatory neurons consistently promotes population synchrony regardless of the population influencing astrocytic dynamics, whereas selective modulation of inhibitory interneurons induces network quiescence via strong suppression. Under dual-target gliotransmission, network synchrony is dictated by the population driving astrocytic dynamics: excitatory-only drive promotes synchrony, while combined or inhibitory-specific drive preserves asynchronous states. Furthermore, the model reveals that astrocytic signaling kinetics provide an additional temporal control mechanism that regulates the frequency and persistence of self sustained up states.

When we appreciate visual art, we often take our time: We visit museums or galleries and deliberately engage with artworks and their meaning. Yet, our liking might at least partly be shaped by incidental encoding of the artworks' visual features, a process initially independent from sustained engagement. Here, we show that liking for visual art can be predicted from rapid and automatic responses in the visual brain. We recorded EEG responses while participants viewed 999 artworks from the Vienna Art Picture System (VAPS), which features complex artworks that span different categories and art styles. Artworks were shown with a 300ms stimulus-onset-asynchrony (150ms on, 150 off), and participants responded to occasional target images. Using representational similarity analysis, we predicted the geometry of neural representations from the artworks' similarity in liking (based on ratings from a different group of observers), while simultaneously controlling for similarities in other properties (category, art style, valence, arousal, and familiarity). Crucially, neural responses emerging within the first 150ms were robustly connected to liking, even though the task did not afford any liking judgments. Strikingly, this early neural correlate of liking was also observed in a second experiment, in which we presented the same 999 artworks with a 50ms stimulus-onset-asynchrony. Our results show that even under ultra-fast presentation conditions, automatic neural responses rapidly indicate liking. While fully appreciating an artwork may still need deliberate and prolonged engagement, the perceptual basis of whether we like or dislike an artwork seems to be laid in a split second.

Age is a fundamental social attribute that affects rights, responsibilities and even social status. Color information of the face can affect perceived age, but the nature of its contribution, independent of luminance information, remains underexplored. To better characterize the role of color in age perception, we conducted two experiments. In our first experiment, we tested whether natural age-related changes in facial color alone can drive changes in perceived age. In our second experiment, we manipulated red-green contrast across faces aged 18-91 and tested the effect of this manipulation of perceived youthfulness. Our findings demonstrate that natural age-related color changes do provide cues for age perception, dissociable from luminance information. We also found that the optimal red-green contrast level to make faces look the youngest was significantly higher than the originals, meaning that faces tended to appear younger when their redness was accentuated, even though average face redness increases with age. Our analysis revealed that younger faces required greater increases in redness level compared to their older counterparts to achieve the youngest appearance. These results suggest that the spatial distribution of redness contributes to apparent youthfulness, especially in younger faces.

Changes in pupil size have been widely used as an indirect measure of attention. However, attention engages on a faster timescale than the pupillary response. Therefore, the slow time course of the pupil response makes it difficult to distinguish attentional responses to stimuli presented in close temporal succession. Although signal deconvolution has been used in previous research to estimate latent attentional responses from the pupil signal, existing methods primarily model dilation without considering the visual stimulus evoked constriction phase of the pupil. Here, we introduce a dual-kernel deconvolution model that models constriction and dilation using separate impulse response functions. Using an RSVP paradigm with emotional faces as target stimuli, we demonstrate that the proposed model estimates latent attentional response despite substantial overlap in the measured pupil signal. Importantly, the recovered attentional response resembles the time course of attentional allocation previously predicted by several models of Attentional Blink. Together, these findings demonstrate that the dynamics of latent attentional episodes can be recovered from the pupil response with a temporal resolution previously obscured by the sluggishness of the pupillary response. The proposed method is also provided as a usable software package that serves as a general tool for investigating the temporal dynamics of attention in rapid visual paradigms.

Financial exploitation often arises in social contexts in which individuals must evaluate fairness, trust, and interpersonal information during financial decisions. We examined whether neural responses to fairness in social versus nonsocial contexts relate to financial exploitation vulnerability and whether these relationships vary with age, emotional intelligence, and affect. A total of 310 adults spanning ages 21-89 completed an fMRI Ultimatum Game in which they accepted or rejected offers from human and computer partners, along with measures of financial exploitation vulnerability, emotional intelligence, and affect. Participants were strongly sensitive to offer fairness, and individual differences in behavioral sensitivity to partner type were associated with social-relative anterior insula responses to unfairness. Greater baseline negative affect was associated with stronger anterior insula differentiation between social and nonsocial fairness. Financial exploitation vulnerability was associated with prefrontal responses to social-relative fairness in ways that depended on individual characteristics: age moderated its association with dorsomedial prefrontal cortex activation, whereas emotional intelligence moderated its association with dorsolateral prefrontal cortex activation. Whole-brain connectivity analyses further showed that age moderated associations between emotional intelligence and executive control network connectivity with posterior cingulate and right temporoparietal regions. These findings suggest that vulnerability to financial exploitation is reflected in how social context is integrated with affective and individual characteristics across adulthood.

Microstate analysis studies dynamic brain activity through a discrete representation of electroencephalographic (EEG) signals and their topography. While the body of literature describing findings in diverse experimental and clinical settings is growing rapidly, the effects of several core components of the microstate algorithm are still insufficiently understood. This study systematically investigates how two common temporal smoothing algorithms, parametric smoothing and GFP peak interpolation, influence a wide range of microstate sequence parameters, and how they shape the integrated picture of brain dynamics inferred from these metrics. We study eyes-closed resting-state EEG recordings from N=201 healthy subjects from a public EEG database. For the elementary temporal metrics, we find that widening smoothing windows markedly increase microstate duration, reduce occurrence rates, but leave coverage mostly unaffected. In terms of simple syntax properties, smoothing windows above 30 ms yield an inconsistent picture with a memoryless, first-order Markov structure at short time scales, but a positive Hurst phenomenon suggestive of long-range correlations at long time scales. Evidence of asymmetry and non-equilibrium dynamics are suppressed by smoothing. At intermediate time scales, microstate periodicity was only observed up to a window size of 28 ms. Smoothing had a strong impact on the complexity metrics entropy rate and excess entropy, decreasing the former and increasing the latter. For most metrics, the effects of GFP peak interpolation were comparable to those of smoothing windows in the range between 28 ms and 44 ms. In conclusion, elementary and higher-order microstate parameters must be interpreted in the context of preprocessing, as many parameters change continuously with smoothing intensity. This limits the comparability of microstate parameters obtained with different preprocessing settings. Strong smoothing with a window size above 30 ms or GFP peak interpolation yields a fundamentally different picture of resting state brain dynamics, i.e. that of a memoryless, aperiodic, time-reversible first-order Markov chain. Since insights from other neurophysiological modalities suggest that time series memory and time-irreversibility are markers of normal brain activity, our recommendation is to avoid intense smoothing to minimize alterations of the initial sequences.

Decision-making requires accumulation of evidence over time, yet exactly how neural circuits implement this computation remains debated. Classical models posit that evidence accumulation is realized by ramping neural activity, whereas recent experimental findings suggest an alternative mechanism in which neural activity remains largely silent before abruptly transitioning into an active decision state. This mechanism leaves a fundamental question unanswered: how is evidence accumulated during the neural activity-silent period? To address this question, we trained excitatory-inhibitory recurrent neural networks (RNNs) with short-term synaptic plasticity (STP) on a reaction-time decision task. We found that after training, the networks learned to leverage STP to make decisions in an activity-silent manner. Choice-tuning analysis revealed a clear functional organization of the network, and connection-pruning analysis uncovered the coupling between choice-selective excitatory and inhibitory neural groups. Guided by the architecture of trained RNNs, we constructed a minimal neural circuit model, in which evidence accumulates in facilitated synaptic efficacy that progressively reshapes the circuit's phase portrait, eventually driving the network state across the stability boundary and triggering a rapid transition to a choice state. We further compared activity before a choice between the theoretical STP and Static models to examine the potential energy benefit of activity-silent decision-making. Together, our results suggest STP as a mechanism for evidence accumulation in activity-silent decision-making and offer insights into the design of energy-efficient artificial neural networks.