NeuroPath: Practically Adopting Motor Imagery Decoding through EEG Signals

Jiani Cao; Kun Wang; Yang Liu; Zhenjiang Li

cs.HCMar 30, 2026

NeuroPath: Practically Adopting Motor Imagery Decoding through EEG Signals

Authors: Jiani Cao, Kun Wang, Yang Liu, Zhenjiang Li

Organizations: Department of Computer Science City University of Hong Kong Hong Kong, China · Department of Computer Science Florida State University Tallahassee, United States

Abstract

Motor Imagery (MI) is an emerging Brain-Computer Interface (BCI) paradigm where a person imagines body movements without physical action. By decoding scalp-recorded electroencephalography (EEG) signals, BCIs establish direct communication to control external devices, offering significant potential in prosthetics, rehabilitation, and human-computer interaction. However, existing solutions remain difficult to deploy. (i) Most employ independent, opaque models for each MI task, lacking a unified architectural foundation. Consequently, these models are trained in isolation, failing to learn robust representations from diverse datasets, resulting in modest performance. (ii) They primarily adopt fixed sensor deployment, whereas real-world setups vary in electrode number and placement, causing models to fail across configurations. (iii) Performance degrades sharply under low-SNR conditions typical of consumer-grade EEG. To address these challenges, we present NeuroPath, a neural architecture for robust MI decoding. NeuroPath takes inspiration from the brain's signal pathway from cortex to scalp, utilizing a deep neural architecture with specialized modules for signal filtering, spatial representation learning, and feature classification, enabling unified decoding. To handle varying electrode configurations, we introduce a spatially aware graph adapter accommodating different electrode numbers and placements. To enhance robustness under low-SNR conditions, NeuroPath incorporates multimodal auxiliary training to refine EEG representations and stabilize performance on noisy real-world data. Evaluations on three consumer-grade and three medical-grade public datasets demonstrate that NeuroPath achieves superior performance.

Figures & tables

Figure 1. The process of MI-based BCI system and its potential applications.

Figure 2. MI signal transmission pathway in the brain.

Figure 3. ERD and ERS patterns during motor imagery of clenching the (a) left hand and (b) right hand.

Figure 4. Overview of the NeuroPath design.

Figure 5. Illustration of the forward generation process.

Figure 6. Brain-inspired MI decoder architecture.

Parameter	Medical	Consumer
Electrode Type	Ag/AgCl Wet	Dry or Saline
Data Resolution	24-bit	16-bit
Voltage Sensitivity	0.39 nV	510 nV
Sampling Rate	High	Low
Channel Density	Dense	Sparse
Price (USD)	$30,000+	$500 -$ 2,000

Table 1. Comparison of key metrics between medical-grade and consumer-grade devices.

Figure 7. (a) International 10-10 electrode system. (b) Task-dependent importance of different channels, based on average signal amplitude in MI tasks.

Figure 8. Cross-dataset pre-training framework integrating heterogeneous datasets through dataset-specific adapters and heads, with a shared MI decoder.

Figure 9. (a) Cross-modality knowledge transfer framework. (b) Skeleton keypoints extracted by MediaPipe.

Dataset	Subjects	Channels	Device	Electrode Type	Resolution (Sensitivity)	Cost (USD)	Collection Time Per User
NeuroPath-DS-32C	12	32	Consumer-grade (Emotiv FLEX 2)	Saline-based	$14\text{\,}\mathrm{b}\mathrm{i}\mathrm{t}$ ( $510\text{\,}\mathrm{nV}$ )	~$1,500	~ $40\text{\,}\mathrm{m}\mathrm{i}\mathrm{n}$
NeuroPath-DS-16C	12	16
NeuroPath-DS-8C	12	8
BCIC-2a ( Brunner et al., 2008 )	9	22	Medical-grade ( e.g. , g.tec)	Gel-based Ag/AgCl	High-precision ( $<$ 1\text{,}\mathrm{\SIUnitSymbolMicro V}$$ )	$>$$30,000$$	~ $58\text{\,}\mathrm{m}\mathrm{i}\mathrm{n}$
BCIC-2b ( Leeb et al., 2008 )	9	3		Gel-based Ag/AgCl		$>$$30,000$$	~ $70\text{\,}\mathrm{m}\mathrm{i}\mathrm{n}$
MI-KU ( Lee et al., 2019 )	19	56	Medical-grade (BrainAmp)	Gel-based Ag/AgCl	$24\text{\,}\mathrm{b}\mathrm{i}\mathrm{t}$ (~ $0.4\text{\,}\mathrm{nV}$ )	$>$$50,000$$	~ $47\text{\,}\mathrm{m}\mathrm{i}\mathrm{n}$

Table 2. Summary of datasets used for evaluation. The first three are self-collected under different channel configurations, while the last three are public benchmark datasets.

Figure 10. Illustration of the experimental setup.

Figure 11. Different electrode numbers and placements for (a) NeuroPath-DS-32C, (b) NeuroPath-DS-16C, (c) NeuroPath-DS-8C, (d) BCIC-2a ( Brunner et al., 2008 ) , (e) BCIC-2b ( Leeb et al., 2008 ) and (f) MI-KU ( Lee et al., 2019 ) .

Figure 12. Data collection procedure of one trial in self-collected NeuroPath-DS. This process is similar to that of public medical-grade datasets, such as BCIC-2a ( Brunner et al., 2008 ) .

Figure 13. Overall performance. “NP” means “NeuroPath”.

Dataset	V1	V2-1	V2-2	V3 (Full)
NeuroPath-DS-32C	67.0%	71.6%	69.3%	73.9%
NeuroPath-DS-16C	66.8%	69.3%	68.7%	72.8%
NeuroPath-DS-8C	66.1%	68.6%	68.5%	69.2%
BCIC-2a	78.1%	82.4%	80.1%	83.7%
BCIC-2b	85.8%	86.3%	85.9%	87.1%
MI-KU	78.4%	85.9%	81.3%	87.0%

Table 3. Ablation study.

ID	Acc(%)	ID	Acc(%)	ID	Acc(%)	ID	Acc(%)	ID	Acc(%)
S01	62.5	S05	64.0	S07	65.5	S08	65.0	S09	57.5
S10	57.0	S11	66.0	S12	68.0	S13	61.5	S14	56.5
S15	66.5	S17	54.5	S19	63.0	S20	71.5	S22	57.5
S23	57.0	S24	60.5	S28	63.0	S31	59.5	S32	71.0
S34	54.0	S37	73.0	S38	59.0	S39	55.0	S40	64.0
S41	59.0	S45	74.0	S46	74.0	S48	63.0	S49	55.5

Table 4. Performance on the BCI illiterate population.

Metric	Value
Model Parameters	60,176
Computational Complexity	352.17 M-Adds
Inference Latency (per trial)	21.22 ms
Power Consumption	145.10 mAh
Battery Impact	3.35% of total battery

Table 5. System overhead of NeuroPath.

Explore similar work

Apr 7, 2026

EEG-MFTNet: An Enhanced EEGNet Architecture with Multi-Scale Temporal Convolutions and Transformer Fusion for Cross-Session Motor Imagery Decoding

Brain-computer interfaces (BCIs) enable direct communication between the brain and external devices, providing critical support for individuals with motor impairments. However, accurate motor imagery (MI) decoding from electroencephalography (EEG) remains challenging due to noise and cross-session variability. This study introduces EEG-MFTNet, a novel deep learning model based on the EEGNet architecture, enhanced with multi-scale temporal convolutions and a Transformer encoder stream. These components are designed to capture both short and long-range temporal dependencies in EEG signals. The model is evaluated on the SHU dataset using a subject-dependent cross-session setup, outperforming baseline models, including EEGNet and its recent derivatives. EEG-MFTNet achieves an average classification accuracy of 58.9% while maintaining low computational complexity and inference latency. The results highlight the model's potential for real-time BCI applications and underscore the importance of architectural innovations in improving MI decoding. This work contributes to the development of more robust and adaptive BCI systems, with implications for assistive technologies and neurorehabilitation.

Panagiotis Andrikopoulos, Siamak Mehrkanoon

Department of Information and Computing Sciences, Utrecht University, Utrecht, Netherlands

MI Classification Brain-Computer Interfaces Hybrid CNN-Transformer Architectures Temporal Convolutional Networks

Ask PDF

Aug 13, 2026

EEG Decoding Using CNN and LSTM Network

Motor imagery (MI) brain--computer interfaces (BCIs) have emerged as a promising approach for establishing flexible communication pathways between the human brain and external devices , particularly for individuals affected by stroke or neurodegenerative disorders. Reliable decoding of motor-imagery electroencephalography (MI-EEG) remains challenging because EEG recordings contain substantial noise and exhibit complex, weakly informative relationships with the underlying brain activity. Although deep learning provides an effective means of learning representations directly from EEG signals, its application to MI-EEG feature learning remains comparatively limited. This study introduces a hybrid deep-learning architecture that integrates a convolutional neural network (CNN) with a bidirectional long short-term memory (bi-LSTM) network. The CNN is used to learn high-level spatial and temporal representations directly from raw MI-EEG recordings, whereas the bi-LSTM models temporal dependencies and relationships among the extracted features. The proposed approach is evaluated using both a publicly available dataset and a privately acquired dataset obtained with an EEG acquisition system. The experimental results indicate that the CNN&bi-LSTM architecture provides robust performance for both two- and three-class motor-imagery classification and demonstrates promising subject-independent decoding capability across the evaluated methods.

Athanasios Karagounis

School of Science, Department of Digital Industry Technologies GR34400, Psachna, Greece

Cross-Subject EEG Decoding MI Classification Brain-Computer Interfaces EEG Decoding

Ask PDF

Sep 26, 2026

EEG-Based Motor Imagery BCI Algorithms and Technologies: A Review

Brain-computer interfaces (BCIs) have emerged as transformative technologies that enable direct communication between the brain and external devices. Among various BCI paradigms, EEG-based motor imagery (MI) has gained prominence due to its simplicity, non-invasiveness, and potential to restore motor function and facilitate rehabilitation for patients with motor impairments. This paper presents a comprehensive review of the most practical processing algorithms developed over the past decade for decoding brain sensorimotor cortex signals. Specifically, this paper discusses the integration of artificial intelligence (AI)-based algorithms, particularly machine learning and deep learning techniques, and their contributions to improving the performance and efficiency of MI-BCI systems in detail. Furthermore, the paper reviews state-of-the-art hardware platforms and emerging converging technologies, including system-on-chip (SoC) architectures, application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), wearable devices, the Internet of Things (IoT), and augmented/virtual reality (AR/VR), and discusses their integration with advanced signal processing algorithms to enable next-generation MI-BCI systems. By highlighting current achievements of EEG-based MI-BCI technology and predicting future research directions that could further enhance real-time capabilities, this paper aims to provide valuable insights for researchers and practitioners, fostering innovation in high-performance EEG-based MI-BCI systems.

Mohammad Hossein Koohi Ghamsari, Seyede Fatemeh Ghamkhari, Siavash Bayat +1

Department of Electrical Engineering, Sharif University of Technology, Azadi St., Tehran, 14588-89694, Iran. · School of Electrical Engineering and Computer Science, KTH Royal Institute of Technology, 100 44 Stockholm, Sweden.

Brain-Computer Interfaces MI Classification EEG Decoding

Ask PDF