脑机接口 / BCI

Hybrid brain-computer interfaces (BCIs) that integrate motor imagery (MI) and steady-state visual evoked potentials (SSVEP) provide high-dimensional neural decoding but typically exceed the computational limits of embedded hardware. To address this, we propose SwitchBraidNet, a compact EEG classification architecture designed for low-power deployment. The model employs a dual-path temporal braid to extract multiscale oscillatory features, an adaptive squeeze-and-excitation spatial switch for electrode gating, and a log-variance readout layer for direct band-power encoding. Furthermore, through systematic quantisation-aware training on the OpenBMI dataset, we compared SwitchBraidNet against four established baselines across FP32, FP16, and INT8 precisions. Experimental results demonstrate superior efficiency and performance, achieving MI accuracy of 69.49% (FP16), SSVEP accuracy of 93.48% (FP32), and a hybrid information transfer rate of 64.82 bits/min (FP16). With an INT8 footprint of only 3.03 KB, SwitchBraidNet maintains high accuracy across varying numerical precisions, demonstrating its suitability for low-power embedded BCI deployment.

Predicate-based computation over state spaces separates a problem specification from the backend that realizes it. Building on the model introduced in arXiv:2606.15027, this paper studies QDSV as a semantic, multi-backend execution framework for quantum-oriented computation. We describe how QDSV, QIntent, and Qruba connect declarative problem intent to a structured semantic representation, realize that representation under heterogeneous backend constraints, and report execution trace outputs that separate model-level semantic outputs from backend-specific observations. The framework supports execution modes that do not require the original problem to be authored as a circuit, while still allowing circuit-compatible artifacts when required. As a case study, we evaluate EEG ictal/interictal classification using prepared signal features from the Bonn and Delhi datasets. The study compares classical machine-learning references, a circuit-first variational quantum classifier baseline, QDSV simulator executions, and controlled IBM Quantum hardware runs. The paper does not claim general quantum advantage or superiority over classical machine learning. Its contribution is a semantic execution validation showing how a problem-first representation can remain stable across simulator and hardware realizations while retaining interpretable execution trace outputs.

The mismatch between continuous acoustic signals and discrete event-driven processing remains a fundamental bottleneck for neuromorphic speech processing. Current systems typically rely on fixed spike encoders, forcing downstream Spiking Neural Networks (SNNs) to compensate for non-adaptive input representations. To address this, we present a learnable residual speech-to-spike encoder jointly trained end-to-end with a Recurrent Leaky Integrate-and-Fire (R-LIF) backbone. We validate this approach on the Google Speech Commands v2 (GSC-v2) benchmark, achieving up to 94.97% accuracy. Notably, the learned encoder remains highly parameter-efficient with a compact 35k-parameter variant that reaches 89.8%, matching or exceeding prior baselines that require an order of magnitude more parameters. Our encoder-focused analysis, including linear probing and gradient-residual inspection, indicates that the encoder does not target faithful signal reconstruction but instead learns task-aligned spike representations that enhance class separability. Finally, we benchmark bio-inspired, hardware-friendly credit assignment by comparing Direct Feedback Alignment (DFA) with surrogate-gradient BPTT under identical architectures and training conditions. We find that DFA reaches 91.5% accuracy, quantifying the performance trade-off of bio-inspired learning rules for modern neuromorphic audio.