Levin Kuhlmann

Kai Sun, Peibo Duan, Yongsheng Huang et al.

Spiking neural networks (SNNs), which are brain-inspired and spike-driven, achieve high energy efficiency. However, a performance gap between SNNs and artificial neural networks (ANNs) still remains. Knowledge distillation (KD) is commonly adopted to improve SNN performance, but existing methods typically enforce uniform alignment across all timesteps, either from a teacher network or through inter-temporal self-distillation, implicitly assuming that per-timestep predictions should be treated equally. In practice, SNN predictions vary and evolve over time, and intermediate timesteps need not all be individually correct even when the final aggregated output is correct. Under such conditions, effective distillation should not force every timestep toward the same supervision target, but instead provide corrective guidance to erroneous timesteps while preserving useful temporal dynamics. To address this issue, we propose Selective Alignment Knowledge Distillation (SeAl-KD), which selectively aligns class-level and temporal knowledge by equalizing competing logits at erroneous timesteps and reweighting temporal alignment based on confidence and inter-timestep similarity. Extensive experiments on static image and neuromorphic event-based datasets demonstrate consistent improvements over existing distillation methods. The code is available at https://github.com/KaiSUN1/SeAl

Abdulaziz Alyahya, Abdallah Al Siyabi, Markus R. Ernst et al.

Continual reinforcement learning challenges agents to acquire new skills while retaining previously learned ones with the goal of improving performance in both past and future tasks. Most existing approaches rely on model-free methods with replay buffers to mitigate catastrophic forgetting; however, these solutions often face significant scalability challenges due to large memory demands. Drawing inspiration from neuroscience, where the brain replays experiences to a predictive World Model rather than directly to the policy, we present ARROW (Augmented Replay for RObust World models), a model-based continual RL algorithm that extends DreamerV3 with a memory-efficient, distribution-matching replay buffer. Unlike standard fixed-size FIFO buffers, ARROW maintains two complementary buffers: a short-term buffer for recent experiences and a long-term buffer that preserves task diversity through intelligent sampling. We evaluate ARROW on two challenging continual RL settings: Tasks without shared structure (Atari), and tasks with shared structure, where knowledge transfer is possible (Procgen CoinRun variants). Compared to model-free and model-based baselines with replay buffers of the same-size, ARROW demonstrates substantially less forgetting on tasks without shared structure, while maintaining comparable forward transfer. Our findings highlight the potential of model-based RL and bio-inspired approaches for continual reinforcement learning, warranting further research.

Guowei Zhang, Yun Zhao, Moein Khajehnejad et al.

Mapping human brain activity to natural images offers a new window into vision and cognition, yet current diffusion-based decoders face a core difficulty: most condition directly on fMRI features without analyzing how visual information is organized across the cortex. This overlooks the brain's hierarchical processing and blurs the roles of early, middle, and late visual areas. We propose Hi-DREAM, a brain-inspired conditional diffusion framework that makes the cortical organization explicit. A region-of-interest (ROI) adapter groups fMRI into early/mid/late streams and converts them into a multi-scale cortical pyramid aligned with the U-Net depth (shallow scales preserve layout and edges; deeper scales emphasize objects and semantics). A lightweight, depth-matched ControlNet injects these scale-specific hints during denoising. The result is an efficient and interpretable decoder in which each signal plays a brain-like role, allowing the model not only to reconstruct images but also to illuminate functional contributions of different visual areas. Experiments on the Natural Scenes Dataset (NSD) show that Hi-DREAM attains state-of-the-art performance on high-level semantic metrics while maintaining competitive low-level fidelity. These findings suggest that structuring conditioning by cortical hierarchy is a powerful alternative to purely data-driven embeddings and provides a useful lens for studying the visual cortex.

Grant Nicholas, Levin Kuhlmann, Gideon Kowadlo

In humans, responsibility for performing a task gradually shifts from the right hemisphere to the left. The Novelty-Routine Hypothesis (NRH) states that the right and left hemispheres are used to perform novel and routine tasks respectively, enabling us to learn a diverse range of novel tasks while performing the task capably. Drawing on the NRH, we develop a reinforcement learning agent with specialised hemispheres that can exploit generalist knowledge from the right-hemisphere to avoid poor initial performance on novel tasks. In addition, we find that this design has minimal impact on its ability to learn novel tasks. We conclude by identifying improvements to our agent and exploring potential expansion to the continual learning setting.

Kai Sun, Peibo Duan, Levin Kuhlmann et al.

The Spiking Neural Network (SNN) has drawn increasing attention for its energy-efficient, event-driven processing and biological plausibility. To train SNNs via backpropagation, surrogate gradients are used to approximate the non-differentiable spike function, but they only maintain nonzero derivatives within a narrow range of membrane potentials near the firing threshold, referred to as the surrogate gradient support width gamma. We identify a major challenge, termed the dilemma of gamma: a relatively large gamma leads to overactivation, characterized by excessive neuron firing, which in turn increases energy consumption, whereas a small gamma causes vanishing gradients and weakens temporal dependencies. To address this, we propose a temporal Inhibitory Leaky Integrate-and-Fire (ILIF) neuron model, inspired by biological inhibitory mechanisms. This model incorporates interconnected inhibitory units for membrane potential and current, effectively mitigating overactivation while preserving gradient propagation. Theoretical analysis demonstrates ILIF effectiveness in overcoming the gamma dilemma, and extensive experiments on multiple datasets show that ILIF improves energy efficiency by reducing firing rates, stabilizes training, and enhances accuracy. The code is available at github.com/kaisun1/ILIF.

Luke Yang, Levin Kuhlmann, Gideon Kowadlo

Continual RL requires an agent to learn new tasks without forgetting previous ones, while improving on both past and future tasks. The most common approaches use model-free algorithms and replay buffers can help to mitigate catastrophic forgetting, but often struggle with scalability due to large memory requirements. Biologically inspired replay suggests replay to a world model, aligning with model-based RL; as opposed to the common setting of replay in model-free algorithms. Model-based RL offers benefits for continual RL by leveraging knowledge of the environment, independent of policy. We introduce WMAR (World Models with Augmented Replay), a model-based RL algorithm with a memory-efficient distribution-matching replay buffer. WMAR extends the well known DreamerV3 algorithm, which employs a simple FIFO buffer and was not tested in continual RL. We evaluated WMAR and DreamerV3, with the same-size replay buffers. They were tested on two scenarios: tasks with shared structure using OpenAI Procgen and tasks without shared structure using the Atari benchmark. WMAR demonstrated favourable properties for continual RL considering metrics for forgetting as well as skill transfer on past and future tasks. Compared to DreamerV3, WMAR showed slight benefits in tasks with shared structure and substantially better forgetting characteristics on tasks without shared structure. Our results suggest that model-based RL with a memory-efficient replay buffer can be an effective approach to continual RL, justifying further research.

Jarrad Rinaldo, Levin Kuhlmann, Jason Friedman et al.

Neural Network movement controllers promise a variety of advantages over conventional control methods, however, they are not widely adopted due to their inability to produce reliably precise movements. This research explores a bilateral neural network architecture as a control system for motor tasks. We aimed to achieve hemispheric specialisation similar to what is observed in humans across different tasks; the dominant system (usually the right hand, left hemisphere) excels at tasks involving coordination and efficiency of movement, and the non-dominant system performs better at tasks requiring positional stability. Specialisation was achieved by training the hemispheres with different loss functions tailored to the expected behaviour of the respective hemispheres. We compared bilateral models with and without specialised hemispheres, with and without inter-hemispheric connectivity (representing the biological Corpus Callosum), and unilateral models with and without specialisation. The models were trained and tested on two tasks common in the human motor control literature: the random reach task, suited to the dominant system, a model with better coordination, and the hold position task, suited to the non-dominant system, a model with more stable movement. Each system outperformed the non-preferred system in its preferred task. For both tasks, a bilateral model outperformed the non-preferred hand and was as good or better than the preferred hand. The results suggest that the hemispheres could collaborate on tasks or work independently to their strengths. This study provides ideas for how a biologically inspired bilateral architecture could be exploited for industrial motor control.

Jens Müller, Hongliu Yang, Matthias Eberlein et al.

Seizure forecasting using machine learning is possible, but the performance is far from ideal, as indicated by many false predictions and low specificity. Here, we examine false and missing alarms of two algorithms on long-term datasets to show that the limitations are less related to classifiers or features, but rather to intrinsic changes in the data. We evaluated two algorithms on three datasets by computing the correlation of false predictions and estimating the information transfer between both classification methods. For 9 out of 12 individuals both methods showed a performance better than chance. For all individuals we observed a positive correlation in predictions. For individuals with strong correlation in false predictions we were able to boost the performance of one method by excluding test samples based on the results of the second method. Substantially different algorithms exhibit a highly consistent performance and a strong coherency in false and missing alarms. Hence, changing the underlying hypothesis of a preictal state of fixed time length prior to each seizure to a proictal state is more helpful than further optimizing classifiers. The outcome is significant for the evaluation of seizure prediction algorithms on continuous data.

Khansa Rasheed, Junaid Qadir, Terence J. O'Brien et al.

Prediction of seizure before they occur is vital for bringing normalcy to the lives of patients. Researchers employed machine learning methods using hand-crafted features for seizure prediction. However, ML methods are too complicated to select the best ML model or best features. Deep Learning methods are beneficial in the sense of automatic feature extraction. One of the roadblocks for accurate seizure prediction is scarcity of epileptic seizure data. This paper addresses this problem by proposing a deep convolutional generative adversarial network to generate synthetic EEG samples. We use two methods to validate synthesized data namely, one-class SVM and a new proposal which we refer to as convolutional epileptic seizure predictor (CESP). Another objective of our study is to evaluate performance of well-known deep learning models (e.g., VGG16, VGG19, ResNet50, and Inceptionv3) by training models on augmented data using transfer learning with average time of 10 min between true prediction and seizure onset. Our results show that CESP model achieves sensitivity of 78.11% and 88.21%, and FPR of 0.27/h and 0.14/h for training on synthesized and testing on real Epilepsyecosystem and CHB-MIT datasets, respectively. Effective results of CESP trained on synthesized data shows that synthetic data acquired the correlation between features and labels very well. We also show that employment of idea of transfer learning and data augmentation in patient-specific manner provides highest accuracy with sensitivity of 90.03% and 0.03 FPR/h which was achieved using Inceptionv3, and that augmenting data with samples generated from DCGAN increased prediction results of our CESP model and Inceptionv3 by 4-5% as compared to state-of-the-art traditional augmentation techniques. Finally, we note that prediction results of CESP achieved by using augmented data are better than chance level for both datasets.

Khansa Rasheed, Adnan Qayyum, Junaid Qadir et al.

With the advancement in artificial intelligence (AI) and machine learning (ML) techniques, researchers are striving towards employing these techniques for advancing clinical practice. One of the key objectives in healthcare is the early detection and prediction of disease to timely provide preventive interventions. This is especially the case for epilepsy, which is characterized by recurrent and unpredictable seizures. Patients can be relieved from the adverse consequences of epileptic seizures if it could somehow be predicted in advance. Despite decades of research, seizure prediction remains an unsolved problem. This is likely to remain at least partly because of the inadequate amount of data to resolve the problem. There have been exciting new developments in ML-based algorithms that have the potential to deliver a paradigm shift in the early and accurate prediction of epileptic seizures. Here we provide a comprehensive review of state-of-the-art ML techniques in early prediction of seizures using EEG signals. We will identify the gaps, challenges, and pitfalls in the current research and recommend future directions.

Ramy Hussein, Mohamed Osama Ahmed, Rabab Ward et al.

Objective: The aim of this study is to develop an efficient and reliable epileptic seizure prediction system using intracranial EEG (iEEG) data, especially for people with drug-resistant epilepsy. The prediction procedure should yield accurate results in a fast enough fashion to alert patients of impending seizures. Methods: We quantitatively analyze the human iEEG data to obtain insights into how the human brain behaves before and between epileptic seizures. We then introduce an efficient pre-processing method for reducing the data size and converting the time-series iEEG data into an image-like format that can be used as inputs to convolutional neural networks (CNNs). Further, we propose a seizure prediction algorithm that uses cooperative multi-scale CNNs for automatic feature learning of iEEG data. Results: 1) iEEG channels contain complementary information and excluding individual channels is not advisable to retain the spatial information needed for accurate prediction of epileptic seizures. 2) The traditional PCA is not a reliable method for iEEG data reduction in seizure prediction. 3) Hand-crafted iEEG features may not be suitable for reliable seizure prediction performance as the iEEG data varies between patients and over time for the same patient. 4) Seizure prediction results show that our algorithm outperforms existing methods by achieving an average sensitivity of 87.85% and AUC score of 0.84. Conclusion: Understanding how the human brain behaves before seizure attacks and far from them facilitates better designs of epileptic seizure predictors. Significance: Accurate seizure prediction algorithms can warn patients about the next seizure attack so they could avoid dangerous activities. Medications could then be administered to abort the impending seizure and minimize the risk of injury.

Matthias Eberlein, Raphael Hildebrand, Ronald Tetzlaff et al.

Epilepsy is the most common neurological disorder and an accurate forecast of seizures would help to overcome the patient's uncertainty and helplessness. In this contribution, we present and discuss a novel methodology for the classification of intracranial electroencephalography (iEEG) for seizure prediction. Contrary to previous approaches, we categorically refrain from an extraction of hand-crafted features and use a convolutional neural network (CNN) topology instead for both the determination of suitable signal characteristics and the binary classification of preictal and interictal segments. Three different models have been evaluated on public datasets with long-term recordings from four dogs and three patients. Overall, our findings demonstrate the general applicability. In this work we discuss the strengths and limitations of our methodology.

Nhan Duy Truong, Levin Kuhlmann, Mohammad Reza Bonyadi et al.

In this article, we propose an approach that can make use of not only labeled EEG signals but also the unlabeled ones which is more accessible. We also suggest the use of data fusion to further improve the seizure prediction accuracy. Data fusion in our vision includes EEG signals, cardiogram signals, body temperature and time. We use the short-time Fourier transform on 28-s EEG windows as a pre-processing step. A generative adversarial network (GAN) is trained in an unsupervised manner where information of seizure onset is disregarded. The trained Discriminator of the GAN is then used as feature extractor. Features generated by the feature extractor are classified by two fully-connected layers (can be replaced by any classifier) for the labeled EEG signals. This semi-supervised seizure prediction method achieves area under the operating characteristic curve (AUC) of 77.68% and 75.47% for the CHBMIT scalp EEG dataset and the Freiburg Hospital intracranial EEG dataset, respectively. Unsupervised training without the need of labeling is important because not only it can be performed in real-time during EEG signal recording, but also it does not require feature engineering effort for each patient.

Nhan Duy Truong, Anh Duy Nguyen, Levin Kuhlmann et al.

Seizure prediction has attracted a growing attention as one of the most challenging predictive data analysis efforts in order to improve the life of patients living with drug-resistant epilepsy and tonic seizures. Many outstanding works have been reporting great results in providing a sensible indirect (warning systems) or direct (interactive neural-stimulation) control over refractory seizures, some of which achieved high performance. However, many works put heavily handcraft feature extraction and/or carefully tailored feature engineering to each patient to achieve very high sensitivity and low false prediction rate for a particular dataset. This limits the benefit of their approaches if a different dataset is used. In this paper we apply Convolutional Neural Networks (CNNs) on different intracranial and scalp electroencephalogram (EEG) datasets and proposed a generalized retrospective and patient-specific seizure prediction method. We use Short-Time Fourier Transform (STFT) on 30-second EEG windows with 50% overlapping to extract information in both frequency and time domains. A standardization step is then applied on STFT components across the whole frequency range to prevent high frequencies features being influenced by those at lower frequencies. A convolutional neural network model is used for both feature extraction and classification to separate preictal segments from interictal ones. The proposed approach achieves sensitivity of 81.4%, 81.2%, 82.3% and false prediction rate (FPR) of 0.06/h, 0.16/h, 0.22/h on Freiburg Hospital intracranial EEG (iEEG) dataset, Children's Hospital of Boston-MIT scalp EEG (sEEG) dataset, and Kaggle American Epilepsy Society Seizure Prediction Challenge's dataset, respectively. Our prediction method is also statistically better than an unspecific random predictor for most of patients in all three datasets.

Nhan Truong, Levin Kuhlmann, Mohammad Reza Bonyadi et al.

Detecting seizure using brain neuroactivations recorded by intracranial electroencephalogram (iEEG) has been widely used for monitoring, diagnosing, and closed-loop therapy of epileptic patients, however, computational efficiency gains are needed if state-of-the-art methods are to be implemented in implanted devices. We present a novel method for automatic seizure detection based on iEEG data that outperforms current state-of-the-art seizure detection methods in terms of computational efficiency while maintaining the accuracy. The proposed algorithm incorporates an automatic channel selection (ACS) engine as a pre-processing stage to the seizure detection procedure. The ACS engine consists of supervised classifiers which aim to find iEEGchannelswhich contribute the most to a seizure. Seizure detection stage involves feature extraction and classification. Feature extraction is performed in both frequency and time domains where spectral power and correlation between channel pairs are calculated. Random Forest is used in classification of interictal, ictal and early ictal periods of iEEG signals. Seizure detection in this paper is retrospective and patient-specific. iEEG data is accessed via Kaggle, provided by International Epilepsy Electro-physiology Portal. The dataset includes a training set of 6.5 hours of interictal data and 41 minin ictal data and a test set of 9.14 hours. Compared to the state-of-the-art on the same dataset, we achieve 49.4% increase in computational efficiency and 400 mins better in average for detection delay. The proposed model is able to detect a seizure onset at 91.95% sensitivity and 94.05% specificity with a mean detection delay of 2.77 s. The area under the curve (AUC) is 96.44%, that is comparable to the current state-of-the-art with AUC of 96.29%.