NEUROBIOLOGY

 

P1. Martyna Bernaciak: Dendritic Spines and Synaptic Plasticity: Structure-Function Perspectives

 

Dendritic spines are small protrusions that extend from the dendritic shaft of many neuron types. They can take various forms that are activity-dependent and can reflect various factors, for instance stage of development of the nervous system or the neuron’s activity. Given their central role in synaptic transmission, there has been a lot of speculation about their precise functions in regulating synaptic strength. This poster aims to show the significance of dendritic spines in synaptic plasticity and address many aspects that can play a key role in dendritic spines activity.

This poster is based on a literature review of recent experimental and theoretical studies on dendritic spine morphology and synaptic plasticity. The articles were selected using databases such as PubMed or Google Scholar, with a focus on studies revolving around the topic of imaging techniques for spine morphology, methods of measuring spine activity, and their role in synaptic signal regulation. The poster outlines questions for future research in this area.

Dendritic spines morphology is usually classified into 3 types: thin, mushroom-shaped and stubby. Additionally, filopodia can also be classified as dendritic spines’ precursors – prominent in the developmental stages of the nervous system, they create space for potential synapses. Those morphologies can only be seen with high-resolution microscopy such as saturated stimulated emission depletion(STED) or expansion microscopy. It has been made clear that cytoskeleton plays a major role in the synaptic plasticity and activity-dependent morphological changes of dendritic spines. Moreover, numerous studies highlight the importance of dendritic spines in voltage compartmentalization and biochemical compartmentalization.

Dendritic spines are vital structures that impact the signal transmission at a dendrite through both voltage and biochemical compartmentalization. They play a key role in synaptic plasticity, through many mechanisms via changes in cytoskeleton, postsynaptic density(PSD) and the regulation of both metabolic and ionic receptors. In regards to dendritic spines research, there is a constant search for different techniques for obtaining both morphology and activity simultaneously, where techniques such as STED or tissue expansion come into practice.

 

 

P2. Wiktoria Podolecka, Jacek Wróbel, Mark Hunt: Damage to the nasal epithelium leads to changes in electrophysiological activity

 

The objective of this study was to investigate how signals originating from the nasal epithelium (NE) influence patterns of electrical activity. Alterations in the sense of smell are increasingly recognized as early signs of several major neurological diseases, many of which are linked to disruptions in normal brain activity. Olfactory sensory neurons (OSNs) transmit information from the NE to the olfactory bulb (OB).

Adult male Wistar rats were surgically implanted with electrodes in the OB, prefrontal cortex (PFC), and ventral striatum (VS), as well as EEG electrodes on the skull. To induce NE damage, one group received intranasal gadolinium, while control animals were given saline. Behavioural alterations were evaluated using a variety of tests. Olfactory function was assessed with the hidden cookie test, accompanied by local field potential recordings. Brain activity during sleep was monitored every four days. NE integrity was analysed by measuring olfactory marker protein (OMP) levels at 5, 15, and 22 days following treatment, and by assessing NE thickness using hematoxylin and eosin staining.

Rats treated with gadolinium exhibited delayed performance in the hidden cookie test and displayed heightened anxiety levels compared to control animals. During wakefulness, local field potentials recorded from the OB showed reduced amplitudes in both the respiration-related frequency band (1–10 Hz) and the gamma range (30–90 Hz) in the treated group. Across all animals, stronger respiratory rhythms were correlated with improved olfactory performance. Notably, slow-wave activity characteristic of classical sleep remained intact. Immunohistochemical analysis confirmed OSN damage in gadolinium-treated rats, as evidenced by reduced expression of OMP.

These results suggest that intranasal gadolinium administration in rats effectively models anosmia. Additionally, the presence of functional olfactory sensory neurons is crucial for preserving normal oscillatory activity in the OB.

 

 

P3. Julia Kosowska: Retrovirus mediated overexpression of Achaete-scute complex-like-1 ehnaces proliferative activity of oligodendrocyte progenitor cells in the postnatal cerebral cortex in vivo

 

The proneural transcription factor Achaete-scute complex-like 1 (Ascl1) plays an important

role in the development of the cerebral cortex and is implicated both in gliogenesis and

neurogenesis. Therefore, Ascl1 has been widely used to reprogram non-neuronal cells into

neurons in vitro. Previous work showed that Ascl1-mediated glia-to-neuron reprogramming

is less efficient in vivo; however, Ascl1 overexpression enhances the proliferative activity of

oligodendrocyte precursor cells (OPCs) at 12 days post injection (dpi). Based on this I decided to examine the effects of Ascl1 overexpression on cell cycle activity in transduced OPCs and astroglia in mouse postnatal cortex at 4 dpi. To examine the effects of Ascl1 overexpression in a wider context I decided to examine the effects of its phospho-site deficient variant (Ascl1SA6) known for efficient glia-to-neuron conversion, alongside Ascl1.

The experiments were conducted in vivo, in a mouse postnatal cerebral cortex; After transduction with Ascl1 the pups were killed, after receiving an injection of 5-ethynyl-2’-deoxyuridine (EdU) to identify proliferating cells 3h prior to killing. Immunohistochemistry was performed on the cortical tissue, followed by imagining with the confocal microscopy, and analysis with Image J and GraphPad Prism.

In line with the previous work, I show that Ascl1 enhances the cell cycle activity of OPCs but not astroglial cells. In contrast to Asc1, Ascl1SA6 does not significantly affect the proliferative activity in either OPCs or astroglial cells. Interestingly, Ascl1SA6-transduced cells display a significant change in morphology as early as 4 dpi, compared to both control and Ascl1-transduced cells.

Overall, the results indicate the role of Ascl1 as a selective enhancer of oligodendrogliogenesis and provide preliminary evidence that supports the role of Ascl1SA6 as an efficient glia-toneuron reprogramming factor. These results contribute to the general knowledge about the role of Ascl1 in the developing brain which in the future may contribute to enabling the development of therapies for neurodegenerative diseases.

 

 

P4. Aleksandra Bramorska, PhD, Mark Jeremy Hunt, PhD: Dynamic Reorganization of Neuronal Activity in the Olfactory Bulb after Subanesthetic Ketamine

Ketamine is a psychoactive compound known to exert complex effects on brain activity. In rodents, it markedly increases the amplitude of high-frequency oscillations (HFOs) across frontal–limbic regions, with the olfactory bulb (OB) emerging as a central hub in this network. Fast oscillations in local field potentials, including gamma (30–100 Hz) and HFOs (130–180 Hz), are believed to temporally coordinate neuronal spiking. To elucidate the mechanisms linking ketamine-induced HFOs to spiking activity, this study investigated how ketamine-driven changes in OB oscillatory dynamics relate to neuronal firing patterns.

LFP data from freely moving rats (N = 6) were acquired from the OB. Five rats were recorded during a 20-minute baseline period and a 30-minute post-injection period following administration of 25 mg/kg ketamine. Additionally, one rat received a saline injection followed 30 minutes later by ketamine.

After ketamine administration, a decrease in both low- and high-gamma power and an increase in HFO power were observed. Modulation analysis indicated that high-gamma activity was largely replaced by HFO, implying a reorganization of oscillatory dynamics. CSD analysis, overlaid on OB histology, revealed distinct spatial patterns for high gamma after ketamine. Peri-stimulus time histogram analysis identified two spike clusters: one exhibiting synchronized activity with gamma oscillations during the baseline and saline conditions, and another lacking such synchronization. Interestingly, spike activity in both clusters became coordinated with HFO following ketamine treatment. Spike-field coherence analysis confirmed increased spike-phase locking after ketamine administration, indicating a close relationship between spiking activity and HFO.

Ketamine reorganizes neuronal activity in the olfactory bulb (OB), inducing a shift from gamma to high-frequency oscillations (HFO). As the OB orchestrates HFO across corticolimbic networks, this OB-centered switch may represent a key mechanism underlying ketamine’s behavioral and neuropsychiatric effects.

 

 

P5. Sandra Romanis, Sylwia Zawiślak, Karolina Nader, Leszek Kaczmarek and Katarzyna Kalita: The impact of Lcn2 deficiency on social behavior in mice maintained on a low-fat experimental diet 

 

Lipocalin-2 (Lcn2) is a small protein involved in the innate immune response. Elevated levels of Lcn2 have been detected in the plasma and adipose tissue of obese individuals and are implicated in appetite regulation via binding to the melanocortin-4 receptor in the hypothalamus. Increased plasma Lcn2 levels have also been observed in obese pregnant women, which may relate to the observed increased risk of neurodevelopmental disorders in their offspring. Although Lcn2 is expressed at low levels in the brain, its expression can be markedly upregulated in response to stress or neuroinflammation. Lcn2 was shown to influence synaptic plasticity, as well as the morphology and function of neuronal and glial cells. Studies in mice have shown that deletion of the Lcn2 gene leads to increased depressive- and anxiety-like behaviors. In humans, deletions and duplications within the genomic region containing Lcn2 have been reported in individuals diagnosed with autism and intellectual disability. Despite these findings, the role of Lipocalin-2 deficiency — particularly in the context of maternal obesity remains unclear. Therefore, the aim of this study is to investigate the effects of maternal diet and Lcn2 gene deletion on social behavior in mice. The research hypothesis assumes that the absence of Lipocalin-2 is associated with social behavior deficits.

This study examined Lcn2 dynamics and behavior in mouse offspring. All pregnant females and offspring were maintained on a low-fat experimental diet (10% kcal from fat). Lcn2 levels in plasma were measured by ELISA at developmental stages (P0–P28) in C57BL/6 offspring. Offspring of Lcn2Het mice (WT and Lcn2KO) were used to assess genotype-related behavioral changes. The social behavior was evaluated using the Eco-HAB system (in-cohort sociability, response to unfamiliar bedding, and general mouse activity) and the Three-Chamber test (response to social olfactory cues).

We showed that plasma levels of Lcn2 change across postnatal development, with the highest concentrations at P0 and P7, followed by a significant decline at later time points (P14, P21, P28). Using the Eco-HAB system, we showed that Lcn2KO mice of both sexes exhibited significantly reduced locomotor activity during the first hour of testing. Lcn2KO mice also demonstrated reduced in-cohort sociability, reaching statistical significance in females but not in males. In response to the odor stimulus, KO males showed a statistically significant increase in preference for the social odor, while KO females displayed a slight, non-significant decrease. These findings were consistent with results from the Three-Chamber test, where KO males again demonstrated a statistically significant increase in the social odor preference index, while KO females showed a marked, though non-significant, decrease in this index

Developmental changes in Lcn2 levels suggest that this protein may play a role in early postnatal brain maturation. Behavioral test results indicate that Lcn2 may be involved in the sex-dependent modulation of social behavior in mice. The observed reduction in exploratory behavior during initial exposure to a novel environment may reflect increased anxiety-like responses in Lcn2 KO animals. These findings highlight a potential role of Lcn2 in neurodevelopmental processes relevant to social and emotional behavior. Further studies are needed to elucidate the role of Lcn2 in regulating offspring behavior in the context of maternal diet.

 

 

P6. Dominique Hahn Bedoya, Ermis Ryakiotakis, Joanna Sadowska, Ewelina Knapska: Histological and behavioral impact of olfactory inhibition through internasal gadolinium administration in mice

 

Our aim is to identify the optimal gadolinium dosage that can effectively deactivate olfaction. We seek to achieve proper nasal epithelium damage to inhibit olfaction while minimizing adverse effects of Gd exposure and avoid behavior impairment. We also want to demonstrate that Gd can be used to induce reversible olfaction inhibition.

Gadolinium (Gd) was administered intranasally to mice under anesthesia at two different volumes (5 μL and 10 μL) and concentrations (30 mg/mL and 60 mg/mL) to determine the most effective dosage for inducing nasal epithelium damage. To evaluate changes in hedonic drive, the sucrose preference test (SPT) was performed before and after Gd administration, measuring the relative consumption of sucrose solution versus water. Mice were habituated to sucrose for three days prior to Gd administration, they were left overnight for recovery, and tested the following days. Nasal epithelium damage and regeneration was assessed after 6 and 13 days post-administration. Epithelium thickness and tissue integrity were assessed in nasal cavity slices stained with hematoxylin and eosin (H&E).

SPTs revealed that Gd administration induced anhedonia in groups that receive a dosage of 10 µL per nostril, independent of Gd concentration, during short-term sucrose exposure. These effects were present up to five days post-administration but were completely reversed when sucrose availability was extended on day 5. Histological evaluation demonstrated that Gd administration produced short-term nasal epithelium damage, which was restored after 13 days.

Our results indicate that intranasal Gd administration at both 30 and 60 mg/mL effectively induces temporal nasal epithelium damage, which is restored within 13 days. In addition, we demonstrate that higher dosages of Gd can negatively affect mouse behavior, although these effects are reduced at lower dosages. This study highlights the potential of Gd as a tool for olfactory manipulation research.

 

 

P7. Taisiia Prosvirova, Wiktoria Podolecka, MSc, Mark Jeremy Hunt, PhD: Ketamine-enhanced high-frequency oscillations (130–180 Hz) in the rat olfactory bulb are dependent on stimulation of kainate receptors

 

Ketamine, an NMDA receptor antagonist, used to treat depression, is known to increase the power of high-frequency oscillations (HFOs; 130-180 Hz) in freely moving rats with the olfactory bulb (OB) a known generator of this activity. Since ketamine increases glutamate release in cortical areas, we hypothesized that activation of non-NMDA receptors (AMPA or kainate (KA)) may underlie the generation of this rhythm. In this study, we 1) determined the role of the olfactory bulb (OB) in the generation of ketamine-enhanced HFO in the rat brain 2) used antagonists to determine the role of AMPA and kainic acid receptors in the generation changes in gamma and HFO in the OB.

Adult male Wistar rats were chronically implanted with electrodes and guides in the OB and electrodes in the ventral striatum (VS), prefrontal cortex (PFC), parietal cortex (ECoG-P), and frontal cortex (ECoG-F). Two experimental designs were used: in first, we recorded baseline local field potentials followed by systemic injection of ketamine (25 mg/kg); in the second one, the baseline recording was followed by the local infusion to OB of chosen experimental substance CNQX/NBQX/UBP310/IEM1925 dihydrobromide (0.5 µg) and then systemic ketamine injection (25 mg/kg). Data was recorded from freely moving animals.

Systemic ketamine produced the strongest increase in HFO in the OB, followed by VS and cortical areas (N=19). In a second study (N=7), the infusion of AMPA/KA antagonists (CNQX or NBQX) to the OB reduced ketamine-enhanced HFO power locally and in the VS, and PFC. In a third group (N=6), infusion of the kainate receptor antagonist UBP310 reduced OB and VS HFO more potently than the AMPA receptor antagonist IEM1925.

Our findings demonstrate that ketamine-enhanced HFO observed in the VS and PFC is dependent on OB activity. Within the OB, HFO generation is driven chiefly by activation of kainate receptors.

 

 

 

 COGNITIVE

 

P8. Maciej Padarz: Neurons selective to emotional stimuli revealed by single-cell recordings in humans

 

Processing emotional stimuli is a core function of the human brain, shaping decisions, behaviors, and social interactions. Despite its relevance for mental health and affective neuroscience, the neuronal mechanisms underlying emotional evaluation remain incompletely understood. We investigated whether neurons in limbic and cortical structures are selectively responsive to emotionally salient stimuli. Specifically, we examined (1) which brain regions contain more emotion-selective neurons, (2) which categories of stimuli elicit the strongest selectivity, and (3) whether neuronal responses are better explained by discrete emotion labels (e.g., fear) or by continuous affective dimensions (e.g., valence).

Here, using a unique opportunity to record single-neuron activity from patients suffering from intractable epilepsy, we gathered neuronal responses to emotional images and words. Stimuli were rated on six scales—valence, arousal, disgust, sadness, happiness, and fear. By choosing continuous emotional dimensions, we aimed to estimate neuronal response magnitude as a function of emotion intensity. Single-neuron activity was isolated from intracanal signal using the OSort algorithm.

The analysis revealed a significant number of emotion-selective neurons in areas such as the amygdala and hippocampus. The distribution of selective neurons varied by brain region and stimulus type. Disgust explained the highest proportion of variance in selective neurons’ responses. PCA-based analyses revealed that many neurons were tuned to dimensions separating specific negative emotional states, rather than to global valence gradients, suggesting differentiated encoding within the negative affect spectrum.

These findings support the notion that emotional information is encoded in a distributed and multidimensional fashion across the brain. Importantly, they indicate that discrete emotional categories and affective dimensions may be supported by distinct neuronal mechanisms. Our results contribute to the understanding of how individual neurons represent complex emotional features and provide a basis for future decoding efforts using neuronal population activity.

 

 

P9. Julia Jakubowska, Jean-Jacques Temprado, Rita Sleimen-Malkoun: Effects of metronomes with different stochastic properties on motor variability and adaptability in bimanual coordination in aging

 

This study aims to investigate 1) how the dynamics of bimanual coordination (BC) change with aging and 2) how synchronizing BC with metronomes with different stochastic properties (“coloured” metronomes) influences motor variability and the system’s adaptive capacities.

In two experimental sessions, healthy young and older adults will perform BC tasks involving anti-phase forearm rotations (engaging non-homologous muscles) at various frequencies under three conditions: following a regular metronome, a metronome infused with white (random) noise in its inter-beat intervals, and one infused with pink (complex) noise. Adaptive capacities will be assessed using measures related to spontaneous transitions to the in-phase pattern, preceded by destabilization of the anti-phase pattern, including transition frequency, transition time, and the number of spontaneous transitions across trials. Motor variability, assessed from the dynamical systems (DST) perspective, will be characterized through fluctuations of the relative phase (RP) between the oscillating forearms, with a particular focus on nonlinear analyses.

Based on the loss-of-complexity theory of aging and prior findings from other coordination tasks, we hypothesize that older adults exhibit weaker long-range temporal correlations in RP variability. Furthermore, we expect that following “coloured” metronomes will not only shift temporal properties of RP towards those of the stimulus (consistently with findings from other motor tasks such as gait or tapping) but also modulate adaptive capacities, with pink-noise metronome enhancing adaptability, particularly in older adults.

Alongside changes in the temporal structure of BC dynamics, the expected findings would support the idea that motor variability may serve as a reliable marker of adaptive capacities. Presented approach will be further evaluated as a potential intervention to enhance motor adaptability in older adults.

 

 

P10. Stanisław Adamczyk, Maja Wójcik, Paweł Orłowski, Paweł Lenartowicz, Justyna Hobot, Anastasia Ruban, Mirosław Wyczesany, Michał Wierzchoń, Michał Bola: Beyond the acute effects: No evidence for persistent neurophysiological reorganization in experienced psychedelic users

 

Contemporary psychedelic neuroscience reports acute neurophysiological changes during psychedelic administration, including reduced oscillatory power, increased entropy, and altered connectivity patterns in large-scale brain networks implicated in self-referential processing, salience detection, and cognitive control. These findings are considered to support influential mechanistic models proposing that lasting neurobiological reorganization drives therapeutic outcomes. However, critical gaps remain in understanding whether the acute effects observed underacute laboratory conditionseffects translate to persistent alterations in naturalistic users.

We conducted a multi-level EEG investigation, including experienced psychedelic users versus matched non-users across two independent datasets. We assessed oscillatory power across five frequency bands, neural signal complexity, and source-localized effective connectivity between Default Mode, Salience, and Central Executive networks during drug-free resting-state conditions.

Contrary to predictions derived from acute administration studies, we found no significant group differences in oscillatory power across frequency bands. Psychedelic users demonstrated reduced rather than increased neural complexity during eyes-open conditions. Connectivity analyses revealed no evidence of predicted network reorganization patterns after correction for multiple comparisons.

These predominantly null findings in ecologically valid samples challenge the assumptions that use of psychedelics leads to persistent neurobiological changes. The discrepancy between controlled laboratory findings and naturalistic populations suggests an emerging need for methodological pluralism and consideration of context-dependency before drawing firm conclusions about mechanisms of psychedelics, their lasting effects, and therapeutic applications.

 

 

P11. Robert Kwaśniak, Dariusz Zapała, Paweł Augustynowicz, Piotr Herbut, Arkadiusz Ziółkowski: Hemodynamic activity of ventral attention and default mode network during Cognitive Motor Interference using mobile fNIRS/DOT

 

Diffuse Optical Tomography (DOT), as a high-density variant of functional near-infrared spectroscopy (fNIRS), shows potential due to its increased spatial resolution, while retaining the ease of use that characterizes optical brain imaging techniques. We used the mobility of this solution to investigate the phenomenon of Cognitive Motor Interference (CMI). It occurs when the performance of one or both tasks deteriorates due to the simultaneous execution of a motor and cognitive dual task. Current methods and available literature do not allow for precise localization of the specific locus of dual-task interference in the brain (Leone et al., 2017). For this purpose the aim of the study was to identify neural differences between single-task (ST) and dual-task (DT) using DOT.

60 participants (31 F) took part in our study, aged 18-76 (M = 26,71; SD = 13,04). Individuals performed the task of multiplying two-digit numbers (MC) and resting (REST) while sitting (single-task) and walking on a treadmill (dual-task). The hemodynamic signal using fNIRS Spectrum C23 (Cortivision sp z o.o.) was recorded from 314 channels using 47 emitters and 32 detectors with distances 10, ~16 and ~36 mm. Photon propagation and image reconstruction was done on the ICBM152 model. Calculation of oxyhemoglobine (HbO) and deoxyhemoglobine (HbR) was done for specific labels from the Schaefer Atlas (Schaefer et al., 2018). 

We found a significant difference in response accuracy F(1, 37) = 4.11; p = 0.05 between ST and DT. Participants performed worse on the DT task (M = 70; SE = 2.39) than on the ST task (M = 74.50; SE = 2.83). We observed a decrease of HbO and increase of HbR F(1, 40) = 14.52; p < 0.001; partial η2 = 0.27, during MC in mPFC (Mdiff = -0.021; pbonf = 0.002). Reversed pattern of activation was observed for REST in mPFC which is a part of default mode network (DMN) (Mdiff = 0.024; pbonf = 0.001). Surprisingly, contrary to our predictions such a pattern were also found for left dPFC F(1,30) = 14.80; p < 0.001; partial η2 = 0.33 and right dPFC F(1,32) = 33.605; p < 0.001; partial η2 = 0.51 which is a part of DMN. Decrease of HbO and increase of HbR was found for MC and reversed pattern has been observed for REST for left dPFC (Mdiff = -0.016; pbonf < 0.001; Mdiff = 0.019; pbonf < 0.001) and right dPFC (Mdiff = -0.025; pbonf < 0.001; Mdiff = 0.028; pbonf < 0.001).

Our findings demonstrate that DT coordination is not dependent upon the exclusive prefrontal area but rather involves the interplay of various specialized information-processing systems. Further analysis using functional connectivity is needed to clarify this.

 

 

P12. Hanna Kapusta, dr hab. Piotr Suffczyński: Virtual Forest, Real Benefits? Heart Rate Variability Evidence

The study examined whether exposure to natural environments, either real or virtual, provides benefits for physiological and psychological relaxation. The main objective was to compare the effects of real forest, virtual forest, real urban, and virtual urban conditions on heart rate variability (HRV) and psychological well-being. It was hypothesized that both real and virtual forest exposure would improve relaxation and well-being compared to urban conditions, with the strongest effects expected in the real forest.

A repeated-measures 2×2 experimental design was used, with 27 healthy participants (18-24 years old) exposed to four conditions: real forest, virtual forest, real urban, and virtual urban. HRV was continuously recorded using a Polar H10 chest strap and Elite HRV application. Psychological well-being was measured with the shortened Psychological Well-Being Scales (PWBS). VR conditions were presented using Meta Quest 3 goggles, with visual, auditory, and olfactory stimuli (forest oil in the virtual forest condition). Each session lasted approximately 15 minutes, during which participants sat quietly and were instructed to focus on and contemplate the surrounding environment. Raw HRV data were preprocessed and analyzed with custom Python scripts, followed by statistical testing using repeated-measures ANOVA.

Exposure to the real forest led to increased HRV indicators (higher RMSSD and HF) and lower LF/HF ratio compared to the real urban condition, reflecting stronger parasympathetic activation and greater relaxation. Virtual forest also showed advantages over virtual urban exposure. Heart rate was lower in forest than in urban environments, regardless of presentation mode. Both real and virtual forest exposure significantly increased psychological well-being, with no difference in effect size between them.

Findings confirm that natural environments enhance relaxation and well-being, with the real forest producing the strongest physiological effects. Virtual forest exposure improved psychological well-being comparably to real nature, but induced weaker physiological changes. VR may serve as a partial substitute for real nature in promoting relaxation, especially for individuals with limited access to natural environments, though contact with real nature remains more effective.

 

 

CLINICAL

 

P13. Martyna Najmrocka, Adam Stanek, dr Kinga Szydłowska: How does Dysfunction of the Glymphatic System Affect the Activation and Function of CD4⁺ T cells in Parkinson’s Disease?

 

To assess whether and how impaired cerebral drainage (glymphatic system and meningeal lymphatic vessels) modulates the activation and function of CD4⁺ lymphocytes in Parkinson’s disease, with an emphasis on the shift in the Th1/Th17↔Treg balance and reactivity to α-synuclein epitopes (PECO: P – PD/iRBD and models; E – indicators of drainage dysfunction: DTI-ALPS, PVS, MLVs, sleep/OSA; C – control groups/higher vs. lower drainage; O – CD4 phenotypes, responses to α-syn, inflammatory markers, and clinical correlates).

A literature review was conducted (PubMed/Scopus, 2015–2025; keywords including “Parkinson disease,” “glymphatic/DTI-ALPS/perivascular spaces,” “meningeal/dural lymphatics,” “CD4/Th1/Th17/Treg,” “alpha-synuclein”). Original studies (imaging, immunological, clinical, and preclinical) and high-quality reviews were included; commentaries without data were excluded.

The identified imaging literature indicates that PD is more commonly associated with reduced DTI-ALPS indices and/or increased perivascular space, which correlates with sleep disturbances and poorer cognitive outcomes. Immunological studies describe a Th1/Th17 advantage, a reduction in Treg and CD4⁺ responses to α-synuclein epitopes presented by MHC II. Experimental data suggest that the weakening of meningeal lymphatic vessels exacerbates neuroinflammation and promotes T cell priming in cervical nodes. We propose an “antigen-drainage-priming” model in which impaired clearance increases antigen load (including α-syn), facilitates presentation, and perpetuates CD4⁺ activation.

Converging evidence supports the hypothesis that cerebral drainage dysfunction may drive pathological CD4⁺ responses and contribute to PD progression. Research combining simultaneous imaging of the glymphatic system/MLVs with CD4⁺ immunoprofiling and interventions aimed at improving drainage (e.g., OSA treatment, AQP4 modulation) and Treg/Th17 balance is a priority.

 

 

P14. Magdalena Szponar, Bartłomiej Gmaj, Wojciech Jernajczyk, Jan Kamiński: Leveraging big data to unveil the influence of psychiatric medications on EEG signals

 

Psychiatric medications are widely prescribed to manage a range of mental health disorders, yet the neurophysiological mechanisms underlying their therapeutic effects remain not fully understood. In this study, we investigated how different classes of psychoactive drugs influence multiple EEG features, including spectral power, functional connectivity, and signal complexity.

We analyzed resting-state EEG data from over 30,000 patients across two psychiatric hospitals in Warsaw, encompassing a wide range of diagnoses and treatments. We compared matched patient groups across 14 medication classes (such as antipsychotics (typical and atypical), different types of antidepressants, benzodiazepines, and anti-epileptic drugs) and against drug-naïve individuals.

We revealed distinct electrophysiological patterns associated with each drug class. The most notable differences were observed in patients taking atypical antipsychotics (with over 45% of features showing significant differences, FDR corrected) and benzodiazepines (BDZ, 35%). BDZ use was associated with increased beta power, enhanced occipital beta connectivity, and reduced frontal beta connectivity. Antipsychotics were linked to reduced signal complexity, increased theta power, and decreased gamma power.

These findings demonstrate that psychiatric medications have a substantial effect on EEG signals, indicating that EEG is a valuable tool for assessing how pharmacological interventions modulate brain dynamics. Pharmaco-EEG studies may also be used for predicting treatment outcomes and have potential for individualized therapy. What is more, EEG-based studies involving medicated patients should control for drug effects between groups, for instance via drug-naïve cohorts, washout periods, or statistically balancing treatment groups.

 

 

P15. poster withdrawn

 

 

 

COMPUTATIONAL

 

P16. Maja Marzec: Exploring EEG Features Structure for Neuroscreening: A Study of Dimensionality Reduction Techniques

 

The study aims to evaluate the effectiveness of dimensionality reduction techniques in optimizing EEG signal classification for neuroscreening. It explores how reduced feature representations can retain interpretability while supporting high classification evaltuation metric (AUC). The research addresses the hypothesis that transformed EEG features can perform comparably to original features in classifying normal versus pathological signals, with added benefits of reduced redundancy and improved generalizability.

An extensive EEG dataset, sourced from 39 hospitals, was preprocessed and segmented into 6-second frames. EEG signals underwent standard preprocessing – filtering, resampling, re-referencing, and amplitude-based artifact rejection. Time-domain and frequency-domain features (power spectral densities, coherence, and covariance) were extracted at both global (recording-level) and local (frame-level) scales. Four data scaling strategies (none, standard, robust, min-max) were applied to investigate their impact on the raw structure of features. Principal Component Analysis (PCA) and Independent Component Analysis (ICA) were used to reduce dimensionality and derive latent features. The structure of the transformed feature space was explored through transformation matrices (loadings/unmixing weights), which revealed dominant feature group contributions. Visual tools such as Pareto and scree plots were used to identify the number of meaningful components, while UMAP was employed to assess class separability and feature topology in reduced spaces. Additionally, an innovative algorithm named Selective Activation was developed to trace component influence from global to local scale, enabling interpretable visualization of EEG activation patterns.

Both PCA and ICA effectively reduced the dimensionality of the EEG feature space while preserving classification performance. The Gradient-Boosted Ensemble model achieved similar AUC scores when trained on reduced features compared to the full feature set, confirming the usefulness of dimensionality reduction. Exploration of transformation matrices showed that dominant feature types varied depending on the applied scaling method. Despite significant reduction, the transformed feature space retained physiologically meaningful structure. UMAP visualizations highlighted clustering by hospital origin but did not show clear separation between normal and pathological signals. The Selective Activation algorithm enabled identification of EEG segments most representative of individual components, improving interpretability and clinical utility. These findings support the development of efficient and explainable EEG-based neuroscreening tools.

Dimensionality reduction can streamline EEG signal classification without sacrificing AUC or interpretability. The study introduces a framework for efficient and explainable EEG-based neuroscreening. The Selective Activation method bridges machine learning models and clinical relevance, supporting the trust and usability of automated diagnostic tools among Health Care Professionals.

 

 

P17. Hanna Moczek, Anna Duszyk-Bogorodzka, Piotr Durka: Estimation of the temporal relationships between slow oscillations and sleep spindles in sleep EEG

 

This study aimed to examine a novel method for the estimation of the relationships between sleep spindles and slow oscillations during typical sleep, in the context of findings reported in the literature regarding the predominance of spindles in up-states of slow oscillations. Both graphoelements are related to memory processes, and there is a theory that their coupling enhances neuroplasticity and memory trace consolidation.

To detect and parametrize slow oscillations and sleep spindles we use a recent implementation of the multivariate matching pursuit (empi), available at https://github.com/develancer/empi, decomposing EEG into a linear sum of Gabor functions (sines with Gaussian envelopes). From these decompositions, we chose the functions corresponding to sleep spindles and slow oscillations based on their frequencies and time lengths, with an additional threshold for the minimal amplitude (12.5 μV p2p for spindles, 120 μV for SWA). Then we chose those pairs of slow waves and spindles, in which the time center of the sleep spindle falls within the time epoch between the maximum and minimum of a co-occurring slow wave. Depending on the temporal width of a given spindle, its oscillations could either be entirely contained within this segment or extend well beyond it—particularly in cases where the spindle center was located near the segment boundary.

Fully automatic detection of sleep spindles, slow oscillations and their co-occurrences allows for robust testing of hypotheses related to the neuroplasticity and memory trace consolidations. A cumulative analysis of 100 sleep recordings from the MASS (The Montreal Archive of Sleep Studies) database revealed that the peak density of fast spindles occurred at the outer edges of the intervals, near the maxima of the slow oscillation. Statistical analysis was conducted for 66 of the 100 recordings, where at least five fast and five slow spindles occurred on each slope of the slow wave. While the results are in general coherent with previously reported effects, further research is needed to determine the dependence of results on the choice of parameters defining the structures and their co-occurences.

We proposed a novel method for a fully parametric investigation of the phasic correlations between slow oscillations and sleep spindles. Preliminary results obtained from recordings of healthy individuals suggest that the method is suitable for further applications.

 

 

P18. Maria Waligórska, Tomasz Wolak: Predicting Brain Age with Machine Learning based on MRI Morphometry: Feature Engineering, Models and Atlas Comparison

Brain aging is a complex and dynamic process influenced by various biological and environmental factors. Assessing the difference between brain-estimated age and actual age could aid in the early detection and diagnosis of age-related brain disorders and help to distinguish between different groups of patients.

Machine learning algorithms can be used to estimate age based on structural brain MRI, as morphological changes in the human brain follow a specific pattern of growth and atrophy throughout development and healthy aging. We explored different models, preprocessing strategies, including dimensionality reduction with Principal Component Analysis, and compared model performances across different anatomical brain atlases. Additionally, we investigated brain structures that have the greatest impact on age prediction. Unlike many studies that focus on markers of dementia, we highlight brain structures that change in a stable and most consistent manner throughout life, rather than those that atrophy most rapidly in old age. We present an analysis using two model explainability methods: Permutation Importance and SHAP.

Our models achieved results comparable to previous research. It is worth noting that our dataset contained the data from patients spanning the entire age spectrum, whereas many studies focus on narrower age ranges, even though estimating brain age is particularly challenging at the extremes. Support Vector Machine and a simple feed-forward network outperformed Random Forest and more complex neural network architectures. We demonstrate that combining carefully selected features from multiple atlases (ASEG and A2009) leads to improved model performance, surpassing models trained on one atlas sets and PCA-transformed data. Permutation Importance indicated that, in the ASEG atlas, the brainstem was the most influential feature contributing to the mean absolute error, followed by total cortical gray matter volume reduction and subcortical regions such as the left accumbens area and left caudate nucleus. The SHAP method highlighted the enlargement of the lateral, third, and fourth ventricles, choroid plexus volume, nucleus accumbens volume. Presence of FreeSurfer’s “non-WM hypointensities” likely capture perivascular spaces and iron or mineral deposits, both of which accumulate with age and are readily observed on structural MRI. In the A2009 atlas, Permutation Importance identified the posterior segment of the lateral fissure (Sylvian fissure), middle occipital gyrus, Superior frontal sulcus regions (bilateral) and frontomarginal gyri as key contributors. SHAP analysis revealed reduced gray matter volume and altered thickness in the inferior circular sulcus of the insula, the parieto-occipital sulcus, and the posterior segment of the lateral fissure, variability in cortical thickness within the calcarine sulcus.

Our results indicate that specific brain regions from FreeSurfer’s ASEG and A2009 atlases are significantly predictive of an individual’s age. Our findings suggest that brain age estimation relies on both subcortical and cortical structural features, with particular importance of regions such as the brainstem, accumbens area, caudate nucleus, and cortical sulci implicated in age-related atrophy. Taken together, these results demonstrate that the model relies on biologically plausible and meaningful features. Furthermore, brain age prediction may not be a well-suited problem for complex neural network architectures.

 

P19. Ewelina Tomana, Nina Härtwich, Reinhard König, Patrick J. C. May, Cezary Sielużycki: What can MEG signals tell us about human auditory cortex connectivity?

 

  Although the human auditory cortex (AC) plays a central role in sound perception, the details of its structural organisation and patterns of connectivity are still unknown. Invasive studies in non-human primates indicate that the AC has a complex, multi-level organisation, yet how these features translate to the human AC is still unclear due to ethical limitations. Non-invasive techniques such as magnetoencephalography (MEG) provide valuable insights into brain activity, though their interpretation is complicated by the complexity of signals arising from millions of neurons. In this work, we build on the hypothesis that the morphology of MEG signals reflects underlying AC connectivity, and use computational modelling with an optimisation algorithm to infer these connectivity patterns.

We adapted the computational AC model by May et al. (2015), which represents core, belt, and parabelt areas of AC as interacting populations of excitatory and inhibitory neurons organised into cortical fields. An evolutionary algorithm was employed to optimise the model’s connectivity patterns by fitting the synthetic signals generated by the model to MEG recordings from healthy human subjects. The resulting optimised connectivity patterns can be interpreted as predictions of in-vivo anatomy of human AC.

Consistent connectivity patterns in the human AC have been observed across two datasets. Our analyses revealed a dominance of feedback over feedforward connections between the thalamus and the core, as well as between the core and the belt, mainly in the left hemisphere. In contrast, feedforward interactions were dominant within the belt area. Decomposition of MEG signals into area-specific contributions showed that event-related fields arise from the combined activity of all auditory cortex areas.

This study demonstrates that computational modelling with parameter estimation can extract subtle, area-specific connectivity features from MEG data, offering a promising tool for investigating human AC organisation. To ensure robustness, future work should also assess alternative optimisation methods to rule out biases from the usage of a single algorithm.

P20. Kalina Nec, Katarzyna Sawicka, Rafał Czajkowski: From Proximity to Dominance: a transparent, species-aware pipeline that converts LMT trajectories  into social, spatial, and dominance indices for cohort comparison.

 

  The goal of this work was to check whether behavioral profiles obtained from Live Mouse Tracker (LMT) – describing social interaction, spatial exploration, and dominance – relate to how well mice learn in the T-maze. 

We wrote a Python pipeline that processes LMT recordings into simple indices for social, spatial, and dominance behaviour. T-maze learning was measured as accuracy across days, and for each mouse we calculated learning slope, mean accuracy, and additional measures such as side bias or early vs late trial performance. Missing days were filled in with estimates from neighbouring sessions. Behavioural indices and learning metrics were matched by RFID and then compared with correlations, regression models, and mixed-effects analysis.

The pipeline works reliably and produces harmonized CSV tables and figures for each animal. We observe clear variability between mice, both in their behavioral indices and in their T-maze learning curves. The framework makes it possible to test how sociality, exploration style, or dominance are connected to learning outcomes. It is flexible enough to adapt to different species or dominance measures, and provides ready-to-use outputs for further statistical analysis or visualization.

By combining LMT and T-maze data in a single workflow, we can better understand how differences in social behavior, exploration, and dominance might shape learning ability. The pipeline is adaptable, transparent, and can serve as a basis for future experiments on the interaction between social context and cognition.

P21. Aleksandra Kołakowska, Joanna Duda-Goławska, Jarosław Żygierewicz, Przemysław Tomalski: Improved Identification of Infant Body Positions Using Machine Learning Approach

Understanding the mechanisms that shape the infants’ cognitive development remains a significant challenge. Research suggests that achieving certain body positions, such as independent sitting or crawling, during infancy serves as a reliable predictor of future motor skills development. By tracking these positions and identifying the infants’ posture, we can also analyse their cognitive progress.

The research expands the state-of-the-art machine learning approach for the identification of infant body positions. The innovative step in the method is the shift from relying on manual annotations to automatic annotation based on the collected sensor data. During the experiments, twelve Inertial Motion Unit (IMU) sensors and three cameras were used to capture the data over four time-points with infants aged from four to twelve months. The sensors were placed on the infants’ arms, legs, head, and torso, and the caregivers’ arms, head, and torso. The measured properties include the angular velocity, acceleration, and the intensity of the magnetic field that are used as features during the classification. There were additional statistical features generated for every sensor. The body posture of an infant was manually annotated using the recorded videos and ELAN software. When extracting the parameters, we used two-second sliding windows with one second overlap between each of the consecutive windows if 75% of the window had the same position label. The aforementioned features and body positions of the infants were included in the dataset used to train the custom CatBoost classifier.
The custom model has a high accuracy when used with overlapping windows, yielding above 85% for each of the position labels. The intended, almost continuous classification, is being manually verified. The preliminary results suggest high accuracy of the model.