doi: 10.1038/srep01319.
A brain-to-brain interface for real-time sharing of sensorimotor information
Affiliations
- PMID: 23448946
- PMCID: PMC3584574
- DOI: 10.1038/srep01319
Item in Clipboard
A brain-to-brain interface for real-time sharing of sensorimotor information
Sci Rep.
2013.
Abstract
A brain-to-brain interface (BTBI) enabled a real-time transfer of behaviorally meaningful sensorimotor information between the brains of two rats. In this BTBI, an "encoder" rat performed sensorimotor tasks that required it to select from two choices of tactile or visual stimuli. While the encoder rat performed the task, samples of its cortical activity were transmitted to matching cortical areas of a "decoder" rat using intracortical microstimulation (ICMS). The decoder rat learned to make similar behavioral selections, guided solely by the information provided by the encoder rat's brain. These results demonstrated that a complex system was formed by coupling the animals' brains, suggesting that BTBIs can enable dyads or networks of animal's brains to exchange, process, and store information and, hence, serve as the basis for studies of novel types of social interaction and for biological computing devices.
Figures
Figure 1. Experimental apparatus scheme of a…
Figure 1. Experimental apparatus scheme of a BTBI for transferring cortical motor signals.
Arrows represent…
Arrows represent the flow of information from the encoder to the decoder rat. In the motor task, the encoder rat has to identify a visual stimulus, signaled by an LED (red circle), and then press one of two levers to receive a small water reward. Meanwhile, M1 neural activity is recorded from the encoder rat and transmitted to the decoder animal, by comparing the pattern of the encoder's M1 to a template trial (previously built with the firing rate average of a trial sample). The difference between the number of spikes in a given trial and the template trial is used to calculate a Zscore. The Zscore is then converted, through a sigmoid function centered on the mean of the template trial, into an ICMS pattern. Thus, the microstimulation patterns varied in real time, according to the number of spikes recorded from the encoder rat's M1, on a trial by trial basis. Once microstimulation is delivered to the M1 cortex of the decoder rat, this animal has to select the same lever pressed by the encoder. Notice that the correct lever to press is cued only by the pattern of the decoder's M1 microstimulation. If the decoder rat pressed the correct lever, both rats were rewarded. Thus, when the information transfer between the brains of the two rats was successful, the encoder rat received an additional reward that served as positive reinforcement.
Figure 2. Behavioral performance using a BTBI…
Figure 2. Behavioral performance using a BTBI for transferring cortical motor signals.
A) Performance of…
A) Performance of encoder and decoder animals during transfer of motor information via a BTBI. The performance of the encoder animals was above 90% in all but one session. The BTBI allowed the decoder animals to repeatedly perform significantly above chance. This performance immediately dropped to chance levels when the cable was disconnected but the system remained fully functional. B) The performance of the decoder animals across a session is presented with a moving average of 10 trials. C) The panel depicts the fraction of right lever presses after different microstimulation patterns were delivered to the decoder's M1. As the number of microstimulation pulses increased, a higher fraction of right lever presses occurred. The microstimulation threshold for response in most animals was situated between 41 and 60 pulses.
Figure 3. Trial examples of a BTBI…
Figure 3. Trial examples of a BTBI for transferring cortical motor signals.
A) Examples of…
A) Examples of M1 neurons recorded while the encoder rat performed the task. Time = 0 corresponds to the lever press. Very different patterns of increased and decreased activity were observed before and after the lever press, suggesting that multiple task parameters were encoded by this M1 ensemble. B) Sample of trial by trial choices of the rat dyad (encoder and decoder) during execution of the motor task. The encoder's performance is depicted by a blue line, while a red line indicates the decoder's choices in the same trials. In trials 4,7,11 and 13 the behavioral response of the decoder rat did not match the one of the encoder. The overall performance of the decoder rat in this session was 69% correct. C) The bars represent the number of encoder's M1 neuronal spikes recorded during each trial. The neuronal ensemble used in this session encoded very accurately each of the behavioral responses. D) Number of ICMS pulses delivered to the decoder's M1 that resulted from the comparison of each trial in C to the template.
Figure 4. Experimental apparatus scheme of a…
Figure 4. Experimental apparatus scheme of a BTBI for transferring cortical tactile information.
A) In…
A) In the tactile discrimination task, the encoder animal was required to sample a variable width aperture using its facial whiskers. The width could be “Narrow” as shown in the left photograph, or “Wide”. After sampling, the encoder animal had to report whether the aperture was narrow or wide by nose poking on a left or right reward port respectively. If correct, the animal received a small water reward. As the encoder explored the aperture, a sample of its S1 activity was recorded, compared with a template trial and then transferred to the decoders' S1 via ICMS. The pattern of microstimulation constantly varied according to the number of spikes recorded from the encoder rat's S1 in each trial. The decoder rat was required to make a response in the reward port corresponding to the width sampled by the encoder, guided only by the microstimulation pattern. If the decoder rat accurately responded in the correct reward port, both rats received a small water reward. Thus, the encoder rat received an additional reward in case both animals of the dyad performed a trial successfully.
Figure 5. Behavioral performance using a brain-to-brain…
Figure 5. Behavioral performance using a brain-to-brain interface to transfer cortical tactile information.
A) Performance…
A) Performance of encoder and decoder animals during operation of a BTBI for tactile information sharing. Notice that the performance of the encoder animals was above 85% in all sessions. The performance of the decoder animals was above 60% in all sessions presented and immediately dropped to chance levels when the cable was disconnected but the system remained fully functional. B) Performance of all decoder animals analyzed with a moving average of 10 trials. C) The panel depicts the fraction of the decoder's responses in the Narrow reward port after different patterns of microstimulation were delivered. As the number of microstimulation pulses increased a higher fraction of responses was observed in the Narrow reward port (Virtual Narrow choice), suggesting that the microstimulation threshold of response for decoder animals was situated between 26-40 pulses.
Figure 6. Trial examples of a BTBI…
Figure 6. Trial examples of a BTBI for transferring cortical tactile signals.
A) Examples of…
A) Examples of S1 neurons recorded while an encoder rat performed the aperture discrimination task. Time = 0 corresponds to the moment the animal breaks the photo beam in front of the discrimination bars. B) Blue lines represent the choices of the encoder rat and red line represents the choices of the decoder rat. In trials 8, 15 and 17 the decoder rat selected the incorrect reward port. C) Number of action potentials recorded from 3 S1 neurons in each trial after the whiskers sampled the discriminanda. Typically, a higher spike count was found for narrow trials, when compared to wide trials. D) Number of pulses delivered to the S1 cortex of the decoder rat in each trial. The number of pulses delivered to the S1 cortex of the decoder rat was directly derived from the number of spikes present in the encoder animal in each trial. The overall performance achieved by the rat dyad in this session was 64% correct trials.
Figure 7. Neural activity in the decoder…
Figure 7. Neural activity in the decoder brain discriminates stimuli applied to the encoder's whiskers.
PSTHs on the left panels show S1 neuronal responses during the wide tactile stimulus whereas PSTHs on the right panels depict narrow tactile stimulus. The top and middle panels show S1 activity recorded in anesthetized encoder and decoder rats while their facial whiskers were passively stimulated by a set of moving bars. The moving bars generate a tactile stimulus exactly like the one produced during the tactile discrimination task. The lower panels represent the decoder rat's S1 activity while receiving ICMS (red traces) via a BTBI that transmitted tactile information from an anesthetized encoder rat which was having its whiskers passively stimulated. Time zero in all panels corresponds either to the tactile stimulus or the last microstimulation pulse. A) A clear peak of S1 activity can be observed immediately after the encoder's whiskers contacted the bars (other peaks occurred due to rebounding of the moving bars). Increased counts of action potentials were typically associated with the narrow stimulus (compare peaks in left versus right panels). B) Like encoder rats, when the decoder rats' whiskers were passively stimulated by the moving bars, clear peaks of S1 activity with different heights can be observed (see left versus right panels). C) When the encoder rats' whiskers were passively stimulated (shown in A) and the BTBI was used to transfer tactile information in real time (shown in C), clear increases in activity were observed in the decoder's S1 cortex after time 0. These S1 firing modulations were larger when the narrow stimulus was applied to the encoders' whiskers when compared to the wide stimulus (see left versus right panels) and were observed in the same S1 neuronal ensembles that responded to natural whisker stimuli (shown in B). Thus, the S1 neuronal responses observed in the decoder rat demonstrate that it learned to use the BTBI and that a representation of the tactile stimuli applied to the encoders' whiskers could be superimposed on the preexisting representation depicting tactile stimuli applied to its own facial whiskers.
Figure 8. Intercontinental brain-to-brain interface to transfer…
Figure 8. Intercontinental brain-to-brain interface to transfer cortical tactile information.
To test the full potentialities…
To test the full potentialities of the BTBI, a brain-to-brain interface to transfer cortical tactile information was established between our laboratory at the IINN-ELS in Brazil and our laboratory at Duke University in the USA. An encoder rat performed a tactile discrimination task at the IINN-ELS. Meanwhile its neuronal activity in S1 was recorded and sent over the internet to our laboratory at Duke University. The sigmoid transformation algorithm was used to transfer the number of action potentials into microstimulation patterns that there were then delivered to the decoder rat's S1 cortex. As the decoder rat made a behavioral response, feedback was sent over the internet to the encoders' chamber back at the IINN-ELS.
References
-
- Hartley R. V. L. Transmission of Information. Bell Technical Journal, 535–564 (1928).
-
- Jackson A. & Zimmermann J. B. Neural interfaces for the brain and spinal cord-restoring motor function. Nat Rev Neurol 8, 690–699 (2012). - PubMed
Publication types
MeSH terms
Grants and funding
LinkOut - more resources
Full Text Sources
Other Literature Sources
