CoChat

BrainNet: A Multi-Person Brain-to-Brain Interface for Direct Collaboration Between Brains

Open access
Jiang, Linxing, Stocco, Andrea, Losey, Darby M., Abernethy, Justin A., Prat, Chantel S., Rao, Rajesh P. N. · 2019 · Scientific Reports

Published version of a preprint

The DOI matches its Crossref record; 30 of the 30 references it cites were found in a registry or an index; 20 of the 29 sources its claims cite were checked for support.

Abstract

We present BrainNet which, to our knowledge, is the first multi-person non-invasive direct brain-to-brain interface for collaborative problem solving. The interface combines electroencephalography (EEG) to record brain signals and transcranial magnetic stimulation (TMS) to deliver information noninvasively to the brain. The interface allows three human subjects to collaborate and solve a task using direct brain-to-brain communication. Two of the three subjects are designated as “Senders” whose brain signals are decoded using real-time EEG data analysis. The decoding process extracts each Sender’s decision about whether to rotate a block in a Tetris-like game before it is dropped to fill a line. The Senders’ decisions are transmitted via the Internet to the brain of a third subject, the “Receiver,” who cannot see the game screen. The Senders’ decisions are delivered to the Receiver’s brain via magnetic stimulation of the occipital cortex. The Receiver integrates the information received from the two Senders and uses an EEG interface to make a decision about either turning the block or keeping it in the same orientation. A second round of the game provides an additional chance for the Senders to evaluate the Receiver’s decision and send feedback to the Receiver’s brain, and for the Receiver to rectify a possible incorrect decision made in the first round. We evaluated the performance of BrainNet in terms of (1) Group-level performance during the game, (2) True/False positive rates of subjects’ decisions, and (3) Mutual information between subjects. Five groups, each with three human subjects, successfully used BrainNet to perform the collaborative task, with an average accuracy of 81.25%. Furthermore, by varying the information reliability of the Senders by artificially injecting noise into one Sender’s signal, we investigated how the Receiver learns to integrate noisy signals in order to make a correct decision. We found that like conventional social networks, BrainNet allows Receivers to learn to trust the Sender who is more reliable, in this case, based solely on the information transmitted directly to their brains. Our results point the way to future brain-to-brain interfaces that enable cooperative problem solving by humans using a “social network” of connected brains.

Overview

CoChat Summary

TL;DR

BrainNet is a non-invasive multi-person brain-to-brain interface combining EEG and TMS, allowing two Senders and one Receiver to collaborate on a Tetris-like task. Five triads achieved 81.25% mean accuracy, and Receivers learned to trust the more reliable Sender based only on brain-delivered information.

Key findings

  • Five triads (15 participants) performed 16 trials; mean accuracy was 0.8125 (13/16 correct), higher than chance by binomial test (p = 0.002), and mean AUC was 0.83 (t(4) = 11.366, p < 0.001; Wilcoxon V = 15, p = 0.031).
  • SSVEP decoding showed correct-frequency power significantly larger than wrong-frequency power during the task (t(15) = 9.709, p < 0.0001 for “Rotate”; t(15) = 10.725, p < 0.0001 for “Do Not Rotate”).
  • Mutual information was MI = 0.336 for good Senders and MI = 0.051 for bad Senders; both were above chance, and the difference was significant (t(8) = 5.187, p = 0.002; Wilcoxon W = 0, p = 0.031). Estimated MI bias was negligible (b = −0.045).
  • Receivers learned Sender reliability: block-by-block beta weights and correlations rose for the good Sender but not the bad Sender, with significant differences in linear trends (beta weight Z = 5.87, p < 0.001; correlation Z = 7.31, p < 0.001).

Methods

  • Design: non-invasive BBI with two Senders and one Receiver; Senders used EEG SSVEP (17 Hz for “Rotate,” 15 Hz for “Do Not Rotate”) to send decisions, and the Receiver received decisions via TMS to left occipital cortex (phosphene for “yes,” no phosphene for “no”) and responded using the same SSVEP interface. Each trial had two rounds, and one Sender was randomly made “bad” in 10 of 16 trials.
  • Sample: 15 healthy participants aged 18–35 (average 22.7 years, eight female), divided into five triads; each session had 16 trials. Trials 8–12 for the first triad were excluded due to a timestamp routine problem.
  • Analysis: group-level accuracy, ROC/AUC, mutual information, and block-by-block regression weights and correlations; statistical tests used t-tests on arcsine-transformed values and Wilcoxon tests.

Limitations

  • Paper-stated: information complexity remained binary with a low bit rate and required disproportionate hardware; the “bad” Sender was purposefully artificial, so natural unreliability from noisy recordings, lack of knowledge, diminished attention, or malicious intent was not tested; the server was task-specific, not general-purpose; uncorrected MI may overestimate shared information, though the estimated bias was negligible.
  • Paper-stated: trials 8–12 for the first triad were excluded due to a timestamp problem; subjects were not explicitly instructed to avoid eye movements, though prior work suggests SSVEP BCIs can be operated without eye movements.
  • Evident from the text: the sample was small (five triads, 16 trials), and the learning analysis concatenated decisions across triads at the group level, limiting individual-level inference; the artificial bad-Sender manipulation limits ecological validity.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1038/s41598-019-41895-7

Cites

30 of 30 references verified0 citation errors · 0 not checkable

Statistics

10 tests recomputed · 0 with the reported and recomputed p on opposite sides of .05

Figures

Figure 1 Architecture of BrainNet. Two participants (“Sender 1” and “Sender 2”) each use a Brain-Computer Interface (BCI) based on EEG to convey information about a collaborative task (here, a Tetris-like game) directly to the brain of the third participant (“Receiver”). Information from each Sender is transmitted over the internet to the Receiver’s brain via a Computer-Brain Interface (CBI) based on TMS. After consciously processing the two inputs from the Senders, the Receiver uses a BCI based on EEG to execute an action in the task. The Senders see the result of this action on their screens (the same updated game state is shown on both screens, as indicated by the red arrow from one Sender’s screen to the other). The Senders then have another opportunity to convey to the Receiver’s brain new information to potentially rectify an incorrect choice in the first round. While our experiment only used two rounds, BrainNet allows an arbitrary number of interactions between the Senders and the Receiver as they collaborate to solve a task. BrainNet differs from a previous interface called “Brainet”12 which combines recordings from multiple monkey brains to perform a common motor task but is unidirectional and does not use stimulation to communicate information back to any of the brains.
Figure 1 Architecture of BrainNet. Two participants (“Sender 1” and “Sender 2”) each use a Brain-Computer Interface (BCI) based on EEG to convey information about a collaborative task (here, a Tetris-like game) directly to the brain of the third participant (“Receiver”). Information from each Sender is transmitted over the internet to the Receiver’s brain via a Computer-Brain Interface (CBI) based on TMS. After consciously processing the two inputs from the Senders, the Receiver uses a BCI based on EEG to execute an action in the task. The Senders see the result of this action on their screens (the same updated game state is shown on both screens, as indicated by the red arrow from one Sender’s screen to the other). The Senders then have another opportunity to convey to the Receiver’s brain new information to potentially rectify an incorrect choice in the first round. While our experiment only used two rounds, BrainNet allows an arbitrary number of interactions between the Senders and the Receiver as they collaborate to solve a task. BrainNet differs from a previous interface called “Brainet”12 which combines recordings from multiple monkey brains to perform a common motor task but is unidirectional and does not use stimulation to communicate information back to any of the brains.CC BY · PubMed Central
Figure 2 Examples of Screens seen by the Receiver and the Senders across Two Rounds. The Receiver sees the three example screens on the left side and the Senders see the screens on the right side. (Top Row) Screens at the beginning of the trial. Note that the Receiver does not see the bottom line with the gap but the Senders do. The Receiver must rely on the Senders to decide whether or not the red block must be rotated to fill the gap and clear the line. (Middle Row) After the Receiver makes a decision in the first round (in this case, “Rotate”), the game state is updated to show the rotated block. (Bottom Row) After the second round, all participants see the results of the Receiver’s action and whether the line was cleared. In this example, the Receiver executed a corrective action to rotate the block again, thereby filling the gap with the bottom part of the block and clearing the line.
Figure 2 Examples of Screens seen by the Receiver and the Senders across Two Rounds. The Receiver sees the three example screens on the left side and the Senders see the screens on the right side. (Top Row) Screens at the beginning of the trial. Note that the Receiver does not see the bottom line with the gap but the Senders do. The Receiver must rely on the Senders to decide whether or not the red block must be rotated to fill the gap and clear the line. (Middle Row) After the Receiver makes a decision in the first round (in this case, “Rotate”), the game state is updated to show the rotated block. (Bottom Row) After the second round, all participants see the results of the Receiver’s action and whether the line was cleared. In this example, the Receiver executed a corrective action to rotate the block again, thereby filling the gap with the bottom part of the block and clearing the line.CC BY · PubMed Central
Figure 3 Performance of Triads of Participants using BrainNet. The plot shows accuracy achieved by each of the five triads of participants. Accuracy was defined as the proportion of correct block rotations achieved by the triad. The dashed line shows the theoretical chance accuracy (0.5).
Figure 3 Performance of Triads of Participants using BrainNet. The plot shows accuracy achieved by each of the five triads of participants. Accuracy was defined as the proportion of correct block rotations achieved by the triad. The dashed line shows the theoretical chance accuracy (0.5).CC BY · PubMed Central
Figure 4 Average Power Spectra of EEG Signals across Subjects during the SSVEP Task. Power values were averaged across one-second epochs and across subjects. The plots show the average power values during the SSVEP task (between dashed lines) and for comparison, the power values three seconds before and after the task. Note that before and after the task, the power values overlap for the two frequencies, whereas during the task, the power of the frequency corresponding to the correct answer is significantly larger.
Figure 4 Average Power Spectra of EEG Signals across Subjects during the SSVEP Task. Power values were averaged across one-second epochs and across subjects. The plots show the average power values during the SSVEP task (between dashed lines) and for comparison, the power values three seconds before and after the task. Note that before and after the task, the power values overlap for the two frequencies, whereas during the task, the power of the frequency corresponding to the correct answer is significantly larger.CC BY · PubMed Central
Figure 5 ROC Curves for the Five Triads of Participants. The plot shows the overall performance of each triad (blue dots) as well as the performances of the two types of Senders (“Good” versus “Bad”) in each triad (green and red dots). See text for details on the experimental design used to create a “Good” versus “Bad” Sender. The superscript on each dot denotes the triad number. Shaded areas represent the area under the curve (AUC) for each triad’s ROC curve. The dashed line denotes chance performance.
Figure 5 ROC Curves for the Five Triads of Participants. The plot shows the overall performance of each triad (blue dots) as well as the performances of the two types of Senders (“Good” versus “Bad”) in each triad (green and red dots). See text for details on the experimental design used to create a “Good” versus “Bad” Sender. The superscript on each dot denotes the triad number. Shaded areas represent the area under the curve (AUC) for each triad’s ROC curve. The dashed line denotes chance performance.CC BY · PubMed Central
Figure 6 Mutual Information transmitted between the Senders and the Receiver. Across all five triads of BrainNet participants, the mutual information transmitted between the Receiver and the “Good” Sender is significantly higher than that between the Receiver and the “Bad” Sender.
Figure 6 Mutual Information transmitted between the Senders and the Receiver. Across all five triads of BrainNet participants, the mutual information transmitted between the Receiver and the “Good” Sender is significantly higher than that between the Receiver and the “Bad” Sender.CC BY · PubMed Central
Figure 7 Quantification of Learning of Sender’s Reliability by Receiver. (Left Panel) Evolution over time of linear regression weights (Beta) for the Receivers’ decision vector and decision vector for each type of Sender for each 4-trial block (see text for details). (Right Panel) Evolution over time of Pearson Correlation Coefficient between the decisions of Receivers and Senders of each type. Both plots exhibit ascending trends for the “Good” Sender but not the “Bad” Sender, suggesting that Receivers learned which Sender was more reliable during the course of their brain-to-brain interactions with the two Senders.
Figure 7 Quantification of Learning of Sender’s Reliability by Receiver. (Left Panel) Evolution over time of linear regression weights (Beta) for the Receivers’ decision vector and decision vector for each type of Sender for each 4-trial block (see text for details). (Right Panel) Evolution over time of Pearson Correlation Coefficient between the decisions of Receivers and Senders of each type. Both plots exhibit ascending trends for the “Good” Sender but not the “Bad” Sender, suggesting that Receivers learned which Sender was more reliable during the course of their brain-to-brain interactions with the two Senders.CC BY · PubMed Central
Figure 8 SSVEP-Based EEG Brain-Computer Interface used by the Senders and the Receiver. Participants conveyed their decisions regarding whether or not to rotate the current block by controlling a cursor (white filled circle) using EEG based steady state visually evoked potentials (SSVEPs). Participants focused on a flashing LED to the left of the screen (depicted as a circle attached to the screen in Fig. 1) to move the cursor leftwards towards the “Yes” side. Focusing on the LED to the right of the screen (flashing at a different frequency) caused the cursor to move rightwards towards the “No” side. If the cursor reached the green “Yes” bar, the interface interpreted the participant’s decision to be rotation of the block (by 180 degrees). If the cursor reached the “No” bar, the interface took the participant’s decision to be to keep the block’s current orientation.
Figure 8 SSVEP-Based EEG Brain-Computer Interface used by the Senders and the Receiver. Participants conveyed their decisions regarding whether or not to rotate the current block by controlling a cursor (white filled circle) using EEG based steady state visually evoked potentials (SSVEPs). Participants focused on a flashing LED to the left of the screen (depicted as a circle attached to the screen in Fig. 1) to move the cursor leftwards towards the “Yes” side. Focusing on the LED to the right of the screen (flashing at a different frequency) caused the cursor to move rightwards towards the “No” side. If the cursor reached the green “Yes” bar, the interface interpreted the participant’s decision to be rotation of the block (by 180 degrees). If the cursor reached the “No” bar, the interface took the participant’s decision to be to keep the block’s current orientation.CC BY · PubMed Central

Verification

Claims the paper makes

16 Valid · 2 Not valid · 4 Not checked · 38 Own results and interpretations

22 more claims · 2 with their cited sources checked

Create a free account to see all 60 claims and every source verdict.

See all claims — free account
  • Probably not a substantial claim. 22
  • It may state two separate claims as one. 20
  • The quote only partly states this claim. 1

Create a free account to see these claims.

Statistics that don’t add up

Each reported p-value recomputed from its test statistic and degrees of freedom.

2 of 2

Statistics recomputed

Every reported test result, recomputed from its test statistic and degrees of freedom. A result shows as not adding up only once an independent check has confirmed it.

  • Doesn’t add upt(4) = -2.897reported p = .021 · recomputed p = .044

    “t(4) = −2.897, p = .021”

  • Doesn’t add upt(8) = 5.187reported p = .002 · recomputed p < .001

    “t(8) = 5.187, p = .002”

  • Consistentt(15) = 9.709reported p < .0001 · recomputed p < .001

    “t(15) = 9.709, p < .0001”

  • Consistentt(15) = 10.725reported p < .0001 · recomputed p < .001

    “t(15) = 10.725, p < .0001”

  • Consistentt(4) = 11.366reported p < .001 · recomputed p < .001

    “t(4) = 11.366, p < .001”

5 of 10 · the rest aren’t loaded

5 more recomputed tests

Create a free account to see every test the Stats check recomputed.

See all tests — free account

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

1 of 1

Citations

Cites

Show all 30 references

Research works printed but not linked: 2

  1. Rao, P. V. et al. Statistical research methods in the life sciences (Duxbury Press, 1998).
  2. Sorrentino, R. & Bianchi, G. Electronic Filter Simulation & Design (McGraw-Hill Education, US, 2007).

Cited by

See all 190

Metadata

Authors
Jiang, L., Stocco, A., Losey, D. M., Abernethy, J. A., Prat, C. S., & Rao, R. P. N.
Affiliations
  • University of Washington (US)
  • Center for the Neural Basis of Cognition (US)
Corresponding author
Rao, Rajesh P. N.
Venue
Scientific Reports
Published
2019-04-16
Volume
9
Issue
1
Pages
6115-6115
ISSN
2045-2322 (ISSN-L)
OpenAlex
W2891977303
Cited by
190
References
27
Citations by year
2026: 10 · 2025: 23 · 2024: 23 · 2023: 20 · 2022: 24 · 2021: 46 · 2020: 28 · 2019: 14
FWCI
16.28
Type
journal article
Language
en
Open access
Licence
Locations
  • Scientific Reports (publishedVersion, cc-by)
  • bioRxiv (Cold Spring Harbor Laboratory) (acceptedVersion, cc-by-nc-nd)
  • Europe PMC (PubMed Central) (publishedVersion, cc-by)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • arXiv (Cornell University) (submittedVersion)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion)
  • arXiv (Cornell University)
Funders
  • Levinson Emerging Scholars Award
  • W. M. Keck Foundation
  • NSF | ENG/OAD | Division of Engineering Education and Centers (1028725)
Topics
  • Life Sciences › Neuroscience › Cognitive Neuroscience › EEG and Brain-Computer Interfaces
  • Life Sciences › Neuroscience › Neurology › Transcranial Magnetic Stimulation Studies
  • Life Sciences › Neuroscience › Cognitive Neuroscience › Neural dynamics and brain function
Keywords
brain-to-brain interfaces, collaborative problem solving, EEG decoding, transcranial magnetic stimulation, decision-making, information integration, trust in information sources, mutual information
MeSH terms
Decision Making, Shared, Adolescent, Adult, Brain, Communication, Cooperative Behavior, Decision Making, Computer-Assisted, Electroencephalography, Female, Humans, Male, Reproducibility of Results, Trust, Transcranial Magnetic Stimulation, Young Adult, Social Networking, Brain-Computer Interfaces, Healthy Volunteers

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

The claim verdicts this report counts are experimental: their accuracy is not yet measured.

A language model drew the claims from the paper on 2026-09-29, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-09-29 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Cited sources are checked with a language model’s help (gpt-6-sol).

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: BrainNet: A Multi-Person Brain-to-Brain Interface for Direct Collaboration Between Brains. https://app.cochat.ai/papers/10.1038/s41598-019-41895-7
The paper
Jiang, L., Stocco, A., Losey, D. M., Abernethy, J. A., Prat, C. S., & Rao, R. P. N. (2019). BrainNet: A Multi-Person Brain-to-Brain Interface for Direct Collaboration Between Brains. Scientific Reports. https://doi.org/10.1038/s41598-019-41895-7

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.