Cross-subject decoding of imagined speech from scalp EEG: a leakage-controlled evaluation on three public corpora

Blessing Isoyiza Adeika; Onyema Osuagwu

eess.SPOct 14, 2026

Cross-subject decoding of imagined speech from scalp EEG: a leakage-controlled evaluation on three public corpora

Authors: Blessing Isoyiza Adeika, Onyema Osuagwu

Organizations: Morgan State University, Baltimore, Maryland, United States of America

Abstract

Objective. A decoder of imagined speech that works for a new user without calibration would make a non-invasive speech interface practical. We tested whether imagined-speech content recorded with scalp electroencephalography (EEG) transfers across people, and examined why accuracy within a single person remains low. Approach. We analyzed three public corpora, ds003626, KaraOne and Track 3 of the 2020 International Brain-Computer Interface Competition, under a protocol that enforces split integrity by assertion and reports exact binomial thresholds, group tests and bootstrap confidence intervals. We trained decoders on each participant and tested them on every other, compared representational geometry across participants without a classifier, applied subject-level alignment, and tested five explanations for the within-subject ceiling. Main results. Across 482 ordered pairs of participants, no pair survived correction for multiple comparisons; mean pair-transfer accuracy lay within 0.006 of chance on every corpus, while every within-subject reference lay above it. Imagery against pre-stimulus rest generalized at about 0.61 after artifact and elapsed-time controls, but its decoder labelled the interval preceding imagery as imagery as often as imagery itself, so the contrast reflects a post-stimulus state, not imagery. Within subject, only the competition corpus decoded above threshold for every participant (0.3147; chance 0.200); syllable count explained little of this, but its frequency profile failed an artifact audit. Significance. In these corpora, imagined-speech content did not transfer between people through any representation we examined. Trial count per class was the only variable that ordered the corpora, which favors larger single-condition datasets over further model development.

Figures & tables

Corpus	Task	Classes	Trials per class	Participants
2020 BCI Competition, Track 3	imagined phrases	5	80	15
ds003626	inner-speech words	4	about 56	10
ds003626	pronounced words	4	about 28	10
KaraOne	imagined words	4	about 12	14
KaraOne	phonemes, imagined and pre-imagery	7	varies	14
KaraOne	binary phonological contrasts	2	varies	14

Table 1: Decoding tasks. Trials per class are per participant. For ds003626 the inner-speech condition is the primary task and the pronounced condition is reported for comparison.

Result	Accuracy	Chance	Null mean	Null 95th pct
KaraOne, imagery vs rest, 5.0 s gap	0.6475	0.500	0.4988	0.5133
KaraOne, imagery vs rest, 2.0 s gap	0.6099	0.500	0.4998	0.5128
KaraOne, rest vs pre-imagery	0.6646	0.500	0.4996	0.5163
KaraOne, pre-imagery vs imagery	0.6122	0.500	0.4999	0.5142
ds003626, three-way condition	0.6065	0.333	0.3317	0.3453
Competition, within subject, band power	0.3147	0.200	0.1969	0.2057

Table 2: Shuffle-label controls for the positive results reported in this paper. Accuracy is the value measured in this study. Labels were permuted within participant and the pipeline re-run 200 times, or 50 times for the network, whose every draw is a full retraining. No null draw reached the observed value in any row, so each empirical

p

equals its minimum of

1/(n+1)

: 0.005 for 200 draws and 0.020 for 50.

Corpus and task	Protocol	Decoder	Accuracy	Macro-F1	Threshold	$n_{\geq}$	$p$ [95% CI]
Competition, 5 phrases	5-fold	band power	0.3273	0.3270	0.235	14/15	0.0003 [0.305, 0.349]
	held out	band power	0.3147	0.3142	0.235	15/15	$<$ 0.0001 [0.294, 0.336]
	LOSO	band power	0.2063	0.1796	0.235	2/15	0.2567 [0.191, 0.222]
ds003626, 4 words, inner	5-fold	band power	0.2595	0.2595	0.305	2/10	0.2695 [0.237, 0.288]
	held out	band power	0.2788	0.2729	0.305	2/10	0.0137 [0.261, 0.300]
	LOSO	band power	0.2590	0.2116	0.305	0/10	0.0947 [0.247, 0.270]

Table 3: Content decoding with the reference decoder. All values were measured in this study; accuracy is the mean over participants. “5-fold” pools sessions and is optimistic; “held out” tests on an unseen session (ds003626) or an unseen block of trials (other corpora). Chance is 0.200 for the five phrases, 0.250 for the four-word tasks and 0.143 for the seven phonemes. Threshold is the exact binomial per-participant threshold at

\alpha=0.05

;

n_{\geq}

is the number of participants at or above it;

p

is the one-sided Wilcoxon test against chance; CI is the bootstrap 95% confidence interval.

Corpus	Protocol	Band	Pairs	Accuracy	Threshold	Pairs surviving FDR [95% CI]
Competition	within	8–13 Hz	10	0.5577	0.569	10/10 [0.538, 0.569]
	within	8–30 Hz	10	0.5853	0.569	10/10 [0.557, 0.600]
	within	30–45 Hz	10	0.6539	0.569	10/10 [0.631, 0.671]
	LOSO	8–13 Hz	10	0.5108	0.569	0/10 [0.503, 0.520]
	LOSO	8–30 Hz	10	0.5107	0.569	1/10 [0.500, 0.519]
	LOSO	30–45 Hz	10	0.5115	0.569	0/10 [0.504, 0.519]

Table 4: Pairwise word separability on matched 128 Hz, 1–45 Hz data. Chance is 0.500. Mean accuracy is over pairs; the CI treats participants as the unit, averaging each participant over pairs first. FDR correction is within each corpus.

Figure 1: Pairwise word separability by corpus, band and protocol. Each box summarizes one value per participant, that participant’s accuracy averaged over the corpus’s word pairs. The box spans the interquartile range, the center line is the median, the diamond is the mean, whiskers extend to 1.5 times the interquartile range and circles mark participants beyond them; fill and hatching identify the band. The dashed line is chance and the dotted segments each corpus’s per-participant threshold.

Figure 2: Every phrase pair of the competition corpus, within subject (top) and across participants (bottom), in four bands. Each box summarizes the 15 participants, drawn as in figure 1 . The dashed line is chance and the dotted line the per-participant threshold (0.569).

Corpus	Alignment	Accuracy	Change	$n_{\geq}$	Wilcoxon $p$
Competition	none	0.2080	–	3/15	0.0956
	Euclidean, inductive 10	0.2063	$-$ 0.0017	2/15	0.2131
	Euclidean, inductive 20	0.2122	$+$ 0.0042	3/15	0.0625
	Euclidean, inductive 40	0.2055	$-$ 0.0025	2/15	0.2534
	Euclidean, transductive	0.2013	$-$ 0.0067	1/15	0.4129
	Riemannian	0.2052	$-$ 0.0028	2/15	0.3661

Table 5: Cross-subject decoding with subject-level alignment (LOSO, band power, 8–30 Hz). Every alignment is fitted per participant from that participant’s unlabeled trials; “inductive

K

” uses only the first

K

trials. Per-participant thresholds are 0.235 (competition), 0.305 (ds003626) and 0.367 (KaraOne). Change is relative to no alignment.

		Pair transfer					Representational similarity
Corpus	Pairs	Within	Transfer	Chance	$\geq$ threshold	FDR	Mean $r$	Null 95th pct	$p$
Competition	210	0.3273	0.1996	0.200	14/210	0/210	$+$ 0.0403	$+$ 0.0601	0.1095
ds003626 inner	90	0.2595	0.2448	0.250	0/90	0/90	$-$ 0.0311	$+$ 0.1176	0.5920
KaraOne	182	0.2986	0.2527	0.250	1/182	0/182	$+$ 0.0179	$+$ 0.1028	0.2985

Table 6: Transfer between individual participants and classifier-free representational similarity. The within-subject reference is 5-fold cross-validation computed in the pair-transfer run; for KaraOne it differs from table 3 in the fourth decimal (0.2986 against 0.2985). Representational similarity is the mean Spearman correlation between participants’ dissimilarity matrices over 105, 45 and 91 participant pairs, built from Riemannian distances between class-mean covariances, against a null that permutes labels within participant (200 permutations).

Figure 3: Accuracy of decoders trained on one participant and tested on another (hatched bars, one value per ordered pair), against each participant’s within-subject accuracy (outlined bars). Both are plotted as proportions so that pairs and participants share an axis. The dashed line is chance and the dotted line the per-participant threshold.

Figure 4: Within-subject accuracy of each participant against the mean accuracy of that participant’s decoder on every other participant. Each point is one participant; the dashed lines mark chance, and

\rho

is Spearman’s rank correlation.

Arm	Rest window	Imagery window	Gap	Accuracy	$n_{\geq}$	Change ( $q$ )
Separated	$-$ 5.0 to $-$ 3.0 s	$+$ 2.0 to $+$ 4.0 s	5.0 s	0.6475	12/14	reference
Adjacent	$-$ 2.0 to 0.0 s	$+$ 2.0 to $+$ 4.0 s	2.0 s	0.6099	11/14	$-$ 0.0377 (0.091)
Late	$-$ 2.0 to 0.0 s	$+$ 5.0 to $+$ 7.0 s	5.0 s	0.5970	12/14	$-$ 0.0505 (0.039)
Adjacent, first half to second			2.0 s	0.5776	5/14	$-$ 0.0699 (0.033)
Adjacent, second half to first			2.0 s	0.6024	10/14	$-$ 0.0452 (0.039)

Table 7: Imagery versus rest on KaraOne under elapsed-time controls (LOSO, re-referenced scalp channels). Windows are relative to stimulus onset. The last column compares each arm with the separated arm;

q

is the Benjamini–Hochberg adjusted

p

-value.

Contrast	Accuracy	Threshold	$n_{\geq}$	$p$ [95% CI]
Rest vs imagery	0.6099	0.552	11/14	0.00006 [0.580, 0.641]
Rest vs pre-imagery	0.6646	0.553	12/14	0.00006 [0.623, 0.706]
Pre-imagery vs imagery	0.6122	0.553	13/14	0.00006 [0.585, 0.647]
Window	Labelled as imagery [95% CI]
Rest	0.394 [0.290, 0.552]
Pre-imagery	0.663 [0.545, 0.762]; above rest in 14/14, $p$ = 0.00006

Table 8: Rest, pre-imagery and imagery windows on KaraOne (LOSO, band power, re-referenced scalp channels). The pre-imagery interval lies between rest and imagery and contains perception of the prompt and preparation to articulate it. Above: each pair of windows decoded directly; threshold is the mean of the per-participant thresholds, and each participant is counted in

n_{\geq}

against their own. Below: fraction of each window type that the rest-versus-imagery decoder, trained on the other participants, labelled as imagery. Each

p

is a one-sided Wilcoxon test across participants; the shuffle-label controls for the upper rows are in table 2 .

Figure 5: Time-resolved decoding with 500 ms windows in 50 ms steps. Each point is the mean accuracy over participants in one window, with its 95% confidence interval; filled points belong to a cluster that survives the cluster-based permutation test over time. The solid line is the accuracy of the full analysis window, the dashed line is chance and the dotted line the per-participant threshold.

Contrast	Prompts	Hold-outs	Accuracy	Threshold	$n_{\geq}$	$p$ ( $q$ )
Nasal	all	28	0.4880	0.532	1/14	0.861 (0.956)
	phonemes	10	0.4980	0.556	0/14	0.725 (0.937)
Bilabial	all	28	0.4984	0.532	2/14	0.749 (0.937)
	phonemes	10	0.5082	0.556	1/14	0.155 (0.311)
/iy/	all	28	0.5883	0.532	12/14	0.0006 (0.0061)
	phonemes	12	0.5117	0.550	4/14	0.377 (0.628)

Table 9: Phonological contrasts on KaraOne with the test prompts held out of training (112 valid hold-out combinations in total). “All” uses all 11 prompts; “phonemes” uses the seven phonemes only.

q

is corrected across the family.

Feature family	Validated	Cells	Largest gain	Accuracy	$p$ ( $q$ )
Connectivity (signed wPLI)	yes	18	$+$ 0.0165	0.2165	0.0345 (0.8824)
Theta–beta coupling	yes	18	$+$ 0.0166	0.2666	0.0176 (0.8824)
Phase at a fixed latency	no	18	$+$ 0.0175	0.1603	0.1446 (0.9761)

Table 10: Content decoding from phase, connectivity and coupling features across three corpora, two region sets and both protocols. Benjamini–Hochberg correction is across all 54 cells.

Task	Protocol	Decoder	Accuracy	Run identifier
Competition, 5 phrases	5-fold	band power	0.3273	20260820_140120_bandpower_lr_bcic2020_within-subject-5fold
Competition, 5 phrases	held out	band power	0.3147	20260821_135506_bandpower_lr_bcic2020_leave-one-block-out
Competition, 5 phrases	LOSO	band power	0.2063	20260820_140122_bandpower_lr_bcic2020_LOSO
ds003626, 4 words, inner	5-fold	band power	0.2595	20260802_002846_bandpower_lr_inner_speech_within-subject-5fold
ds003626, 4 words, inner	held out	band power	0.2786	20260821_151002_bandpower_lr_inner_speech_LOSSO-(leave-one-session-out)
ds003626, 4 words, inner	LOSO	band power	0.2590	20260802_002905_bandpower_lr_inner_speech_LOSO

Table S1a: Run identifiers for the content cells of table 3 of the main text. Accuracy is repeated so that rows can be matched.

Corpus	Alignment	Accuracy	Run identifier
Competition	none	0.2080	20260820_161905_bandpower_lr_bcic2020_LOSO
Competition	Euclidean, inductive 10	0.2063	20260820_162013_bandpower_lr_bcic2020_LOSO
Competition	Euclidean, inductive 20	0.2122	20260820_162038_bandpower_lr_bcic2020_LOSO
Competition	Euclidean, inductive 40	0.2055	20260820_162103_bandpower_lr_bcic2020_LOSO
Competition	Euclidean, transductive	0.2013	20260820_161933_bandpower_lr_bcic2020_LOSO
Competition	Riemannian	0.2052	20260820_162202_bandpower_lr_bcic2020_LOSO

Table S1b: Run identifiers for the alignment cells of table 5 of the main text.

Training set	Rows per fold	Multiplier	Accuracy	Change	$n_{\geq}$	$q$
inner only	149	1.00	0.2788	reference	2/10	n/a
+ pronounced	224	1.50	0.2765	$-$ 0.0023	2/10	1.000
+ visualized	301	2.02	0.2672	$-$ 0.0116	1/10	0.463
+ both	376	2.52	0.2705	$-$ 0.0083	0/10	0.648

Table S2: Inner-speech accuracy on ds003626 (held-out session) when pronounced and visualized trials of the same words are added to training.

q

is from a paired per-participant test with Benjamini–Hochberg correction.

Contrast	Positive class	Negative class	Difference
/iy/	100%	42.9%	57.1
nasal	50.0%	71.4%	21.4
bilabial	50.0%	71.4%	21.4
/uw/	50.0%	66.7%	16.7

Table S3: Proportion of phoneme prompts (as opposed to word prompts) in each class of the KaraOne contrasts, and the difference in percentage points.

Task	Protocol	Decoder	Accuracy	Balanced	Macro-F1	Perm. $p$
Competition, 5 phrases	5-fold	band power	0.3273	0.3273	0.3270	0.0001
Competition, 5 phrases	held out	band power	0.3147	0.3147	0.3142	0.0001
Competition, 5 phrases	LOSO	band power	0.2063	0.2063	0.1796	0.2234
ds003626, 4 words, inner	5-fold	band power	0.2595	0.2595	0.2595	0.2439
ds003626, 4 words, inner	held out	band power	0.2786	0.2788	0.2729	0.0127
ds003626, 4 words, inner	LOSO	band power	0.2590	0.2590	0.2116	0.0898

Table S4: Content cells of table 3 of the main text under class-balanced metrics and an exact permutation test.

Corpus	Features	5-fold	Held out	Pooled minus held out
Competition	six-band bank	0.3273	0.2952	$+$ 0.0322
Competition	8–30 Hz	0.2980	0.2683	$+$ 0.0297
ds003626 inner	six-band bank	0.2595	0.2788	$-$ 0.0193
ds003626 inner	8–30 Hz	0.2537	0.2513	$+$ 0.0024
KaraOne	six-band bank	0.2985	0.2583	$+$ 0.0402
KaraOne	8–30 Hz	0.2622	0.2500	$+$ 0.0122

Table S5: Within-subject accuracy with sessions pooled (5-fold) and with a session, run or block held out, for the two band-power feature sets.

Protocol	Band	Matched	Mismatched	Difference	$p$ ( $q$ )	Matched $p$	Matched 95% CI
within	8–13 Hz	0.5483	0.5634	$+$ 0.0151	0.019 (0.025)	0.0007	[0.527, 0.561]
within	8–30 Hz	0.5731	0.5932	$+$ 0.0201	0.008 (0.025)	0.0004	[0.550, 0.589]
within	62–100 Hz	0.6673	0.6827	$+$ 0.0154	0.084 (0.084)	0.0003	[0.648, 0.686]
within	8–30 + 62–100 Hz	0.6566	0.6785	$+$ 0.0220	0.013 (0.025)	$<$ 0.0001	[0.636, 0.673]
LOSO	8–13 Hz	0.5051	0.5156	$+$ 0.0105	0.115	0.165	[0.494, 0.517]
LOSO	8–30 Hz	0.5063	0.5144	$+$ 0.0082	0.211	0.266	[0.494, 0.519]

Table S6: Pairwise separability on the competition corpus for phrase pairs matched and mismatched in syllable count. Chance is 0.500. “Within” is 5-fold cross-validation within subject. The last two columns test the matched pairs alone against chance.

Contrast	Accuracy	Run identifier
Rest vs imagery	0.6099	20260930_175026_bandpower_lr_karaone_LOSO
Rest vs pre-imagery	0.6646	20260930_182642_bandpower_lr_karaone_LOSO
Pre-imagery vs imagery	0.6122	20260930_190302_bandpower_lr_karaone_LOSO

Table S7: Run identifiers for the three contrasts of table 8 of the main text. Each run carries its own shuffle-label control of 200 draws. The transfer rates in the lower half of that table were computed on the same participant folds as the rest-versus-imagery run.

Cross-Subject EEG Decoding

Cross-subject EEG decoding develops methods that classify or predict brain activity for people not seen during training.

See all 71 papers

Decoding Imagined Speech: A Strictly Subject-Independent Approach Using EEG
Sep 24, 2026Frederik Møllskov Trier, Xiaopeng Mao, Sadasivan PuthusserypadyCross-Subject EEG Decoding Electroencephalography
Department of Health Technology Technical University of Denmark Ørsteds Plads, 2800 Kgs. Lyngby, Denmark
Zero-Shot Imagined Speech Decoding via Imagined-to-Listened MEG Mapping
May 8, 2026Maryam Maghsoudi, Shihab ShammaBrain-Computer Interfaces Neural Decoding
University of Maryland · College Park, MD 20740
Semantic Alignment of EEG and Text: A Contrastive Learning Framework for Decoding Imagined Speech
Oct 14, 2026Shahryar G. Namdari, Francesco Iacomi, Riccardo Barbieri +1EEG Decoding EEG-to-Text Decoding
Department of Electronics, Information and Bioengineering, Politecnico di Milano, Milano, 20133, Italy

Sep 24, 2026J/K move · Enter open · S save

Decoding Imagined Speech: A Strictly Subject-Independent Approach Using EEG

Frederik Møllskov Trier, Xiaopeng Mao, Sadasivan Puthusserypady

Department of Health Technology Technical University of Denmark Ørsteds Plads, 2800 Kgs. Lyngby, Denmark

Cross-Subject EEG Decoding Electroencephalography Brain-Computer Interfaces EEG Decoding

Open paper AskPDF arXiv

Imagined speech decoding from electroencephalography (EEG) has gained increasing attention as a potential communication pathway for individuals with severe motor impairments, yet reported performance often relies on evaluation protocols that do not clearly reflect cross-subject generalization. This study presents a transparent baseline investigation of a multi-class imagined speech EEG dataset under a strictly subject-independent evaluation framework. Two preprocessing and feature extraction pipelines were compared: a time-domain statistical feature approach and a frequency-domain spectral bandpower approach, evaluated using subject-wise cross-validation and trial-level majority voting with a random forest classifier. The spectral pipeline achieved a significantly higher mean trial-wise accuracy than the statistical pipeline (49.03

\pm

4.18% vs. 37.97

\pm

3.79%) for coarse-level classification across subjects. Forward feature selection further indicated that a limited subset of frequency bands captured most of the discriminative information. Overall, this work provides a strong basis for future brain-computer interface studies targeting improved cross-subject generalization in EEG-based imagined speech decoding.

Abstract

Cross-Subject EEG Decoding

Explore similar work

Decoding Imagined Speech: A Strictly Subject-Independent Approach Using EEG

Zero-Shot Imagined Speech Decoding via Imagined-to-Listened MEG Mapping

Semantic Alignment of EEG and Text: A Contrastive Learning Framework for Decoding Imagined Speech