The growing number of depressive people and the overload in primary care services make it necessary to identify depressive states with easily accessible biomarkers such as mobile electroencephalography (EEG). Some studies have addressed this issue by collecting and analyzing EEG resting state in a search of appropriate features and classification methods. Traditionally, EEG resting state classification methods for depression were mainly based on linear or a combination of linear and non-linear features. We hypothesize that participants with ongoing depressive states differ from controls in complex patterns of brain dynamics that can be captured in EEG resting state data, using only nonlinear measures on a few electrodes, making it possible to develop cheap and wearable devices that could be even monitored through smartphones. To validate such a perspective, a resting-state EEG study was conducted with 50 participants, half with depressive state (DEP) and half controls (CTL). A data-driven approach was applied to select the most appropriate time window and electrodes for the EEG analyses, as suggested by Giacometti, as well as the most efficient nonlinear features and classifiers, to distinguish between CTL and DEP participants. Nonlinear features showing temporo-spatial and spectral complexity were selected. The results confirmed that computing nonlinear features from a few selected electrodes in a 15 s time window are sufficient to classify DEP and CTL participants accurately. Finally, after training and testing internally the classifier, the trained machine was applied to EEG resting state data (CTL and DEP) from a publicly available database, validating the capacity of generalization of the classifier with data from different equipment, population, and environment obtaining an accuracy near 100%.

The current study examined how native language (L1) orthographic features influence bilinguals' ability to rapidly form new representations for second-language (L2) written word forms. We assessed eye movements, neural activity, and behavioral performance in sequential bilinguals with L1-L2 orthographical pairs within (Spanish-English) and across (Mandarin Chinese-English) writing systems. Participants were trained on novel English words embedded in naturalistic sentence contexts, followed by a two-day post-training session. Analyses of eye-movement variables, fixation-related potentials, and learning outcomes revealed that Spanish-English bilinguals exhibited more automatic lexical encoding of novel words, reflected in shorter fixations and attenuated early neural responses, whereas Chinese-English bilinguals maintained prolonged attentional engagement across exposures. These divergent exposure-related dynamics predicted post-training performance, with Spanish bilinguals showing greater gains after consolidation. The findings suggest that the neuroplasticity underlying L2 word learning is shaped by pre-existing cortical circuits for native linguistic features, leading to different neural strategies for forming new lexical representations.

ERP studies of the Stroop task typically rely on peak-based measures, which may obscure the complex dynamical structure of neural responses. We propose a Dual Damped Sine (DDS) framework to obtain a low-dimensional, parametric characterization of stimulus-locked activity. Using a public Stroop dataset, DDS parameters were estimated within fixed N200/N450 windows, showing high goodness-of-fit. We then extended this to a sliding-window approach (100 ms) to track time-resolved dynamics at electrode Fz and a frontal ROI. To manage parameter dependence and multiple comparisons, we employed an omnibus multivariate framework (Hotelling's $T^2$) with cluster-based permutation testing. While temporally localized modulations emerged in the late post-stimulus interval, they did not reach significance after correction, suggesting a descriptive rather than confirmatory interpretation for this dataset. Complementary simulations confirmed that key parameters like dominant frequency and damping are reliably recoverable under realistic noise. Overall, DDS modeling provides a structured, exploratory framework for characterizing the temporal organization of cognitive control dynamics.

Emotional ambivalence-the simultaneous activation of conflicting affective states such as desire and guilt, autonomy and attachment, love and fear-is a transdiagnostic feature of persistent psychological distress. Most interventions implicitly push toward resolving such conflict in favour of one pole. We propose an alternative therapeutic target: the integration of ambivalence, whereby contradiction becomes habitable rather than eliminated. We introduce and theoretically ground the musical chimera: a personalized, co-created symbolic-musical object designed to hold two opposing affective poles at once. Drawing together evidence from four literatures-predictive processing of music, the neuroscience of musical reward and its individual differences, large-scale brain networks supporting the narrative self, and the transitional-object tradition-we articulate a single mechanistic model in which the chimera operates as a highlevel generative model that offers an integrated representation where there was split, is rendered rewarding and selfrelevant through personalization, and is consolidated through autonomous repeated listening as a digital transitional object. We derive a set of falsifiable behavioural and neural predictions, delimit the construct from adjacent approaches (therapeutic songwriting, receptive music therapy, and reactive AI-generated music), and specify the bounded role of generative AI as a production tool rather than a therapeutic agent. The account is explicitly prevalidated and hypothesis-generating; we outline the empirical programme required to test it.

BACKGROUND Long-standing intrapsychic conflicts often arise from apparently irreconcilable tensions, such as desire versus affection or autonomy versus dependence. Traditional approaches in psychotherapy describe defense mechanisms or splitting to cope with such conflicts. However, less attention has been given to creative integrative processes that may reconcile opposing tendencies. OBJECTIVE This paper introduces the concept of AI-facilitated symbolic juxtaposition, where generative models are used to create “digital chimeras”—hybrid symbolic constructions integrating objects of desire with affective attributes. We aim to provide a theoretical foundation, operational hypotheses, and clinical protocols for testing this novel framework. METHODS Drawing from psychoanalytic theory (Winnicott’s transitional objects), predictive processing, and neuroscience of the default mode and mentalizing networks, we propose a neuro-symbolic model for symbolic integration. We outline four testable hypotheses: (1) neural integration (DMN coherence), (2) symbolic flexibility, (3) enhancement of attachment security, and (4) accelerated therapeutic outcomes. Empirical validation methods include fMRI, EEG coherence, eye-tracking, attachment interviews, and cognitive flexibility tasks. We also present a clinical implementation protocol with AI-assisted symbolic generation, immersive VR/AR environments, and ethical safeguards. RESULTS As a conceptual and methodological paper, results are presented as expected outcomes. We anticipate that AI-facilitated chimera formation will (a) improve DMN connectivity, (b) enhance cognitive flexibility, (c) increase attachment security, and (d) reduce the number of sessions required for clinically significant change. Clinical protocols emphasize therapist training, patient safety, cultural adaptation, and preservation of therapeutic alliance. CONCLUSIONS AI-facilitated symbolic juxtaposition represents a novel approach to psychotherapy, offering a scientifically grounded and clinically feasible method for resolving long-term intrapsychic conflicts. By combining neuro-symbolic AI, neuroscience, and psychotherapy theory, this framework contributes to the field of digital mental health and sets the stage for future empirical validation across cultural contexts.