The technical premise of ICMS is deceptively simple. By passing carefully shaped pulses of electrical current through microelectrodes positioned in cortical tissue, researchers can evoke neural activity in targeted populations of neurons without requiring any natural sensory input. The timing, amplitude, frequency and spatial pattern of those pulses determine what the brain perceives. In a functioning brain-computer interface, this creates the possibility of a true two-way channel: the device decodes neural signals to infer intent, and it writes information back into the cortex through stimulation. The Tianjin University review, authored by Pengfei Hu, Chong Chen, Yunliang Zang, Xiaohong Li and Dong Ming, organizes the field around this delivery-and-parameter framework, tracing how pulse design and electrode placement jointly shape both perception and long-term circuit change.

The most direct and best-documented application is artificial touch. When electrodes stimulate the primary somatosensory cortex, human participants report localized sensations described as touch, pressure or tingling at specific points on the body, even though no peripheral nerve is involved. Early experiments established that individual electrodes reliably produce perceptible, place-specific sensations. More recent work has pushed further, using multiple electrodes and engineered spatiotemporal patterns to convey richer tactile information, including edges, curvature and even apparent motion across the skin. Human studies have shown that such patterned stimulation improves the controllability and internal structure of artificial touch, allowing users to discriminate features that single-pulse stimulation cannot convey. Yet the review is candid about the gap that remains: synthetic tactile signals still do not reproduce the complexity of natural touch, which combines thousands of mechanoreceptors firing with precisely coordinated timing.

Vision follows a parallel logic. Electrical stimulation of the primary visual cortex produces phosphenes, the perceived spots or lines of light that appear even in people who have been blind for years. Because phosphenes can be evoked across a grid of electrodes, coordinated stimulation of multiple sites can be composed into recognizable shapes and letters, much as individual pixels combine into an image. Experiments with blind participants have demonstrated simple two-dimensional visual patterns and basic object-localization tasks, offering proof of principle that cortical prostheses can deliver usable visual information. Still, the researchers note that current visual prosthesis studies remain confined to relatively simple shapes, letters and localization. Predicting how a given stimulation will be perceived, identifying which electrode combinations produce the most useful phosphenes, and maintaining stable stimulation over long implantation periods remain the field’s most stubborn engineering challenges.

One of the most consequential findings the review synthesizes is that the brain can learn. Animals trained with intracortical stimulation learn to interpret artificial patterns and use them to guide behavior, even when those patterns bear no direct resemblance to natural sensory codes. This observation suggests that the nervous system does not demand an exact mimicry of biology; it can assign meaning to an entirely synthetic neural signal through experience. The implications for future bidirectional brain-computer interfaces are substantial. Rather than painstakingly reverse-engineering natural sensory encoding, engineers may be able to design stimulation schemes that are simpler, more robust and more flexible, and trust cortical plasticity to do the interpretive work. The brain, in effect, becomes a co-designer of the interface.

The review then examines a more ambitious possibility: using ICMS not merely to create momentary sensations but to change how neural circuits function over longer periods. Repeated or precisely timed stimulation can induce plasticity-like changes in cortical networks, the same class of modifications through which the brain normally stores skills and recovers from injury. The authors discuss studies in which paired or activity-dependent stimulation altered functional connectivity between cortical regions. In one particularly striking closed-loop paradigm, spontaneous neural activity recorded from the motor cortex was used to trigger stimulation of the somatosensory cortex with a controlled delay. Because the stimulation arrived at a biologically meaningful time relative to the spontaneous activity, the temporally matched pairing strengthened intercortical coupling and was associated with improved motor recovery in a rat model of brain injury. The result points toward stimulation therapies that reshape circuits rather than simply activating them.

Despite these advances, the authors are explicit that such applications remain largely experimental. Reliable biomarkers that confirm a circuit has actually changed, reproducible stimulation parameters that work across individuals, implantation procedures that are safe over years, and therapeutic benefits that endure all require further validation before any clinical translation. The history of neuromodulation is littered with promising animal results that failed to survive the transition to human trials, and ICMS researchers are aware that plasticity induction, in particular, is exquisitely sensitive to timing, dosage and the state of the tissue being stimulated. What works in a healthy rat motor-sensory loop may behave very differently in an injured or aged human cortex.

A parallel challenge is the hardware itself. Conventional microwire and silicon electrodes are far stiffer than the soft, delicate tissue of the brain, and this mechanical mismatch has consequences. Micromotion between implant and tissue causes damage, inflammation and glial scar formation, which progressively isolate the electrode from the neurons it needs to reach and degrade signal quality over time. Flexible electrodes soften the mechanical mismatch, and the review describes the field’s migration from rigid probes toward flexible, biomimetic designs. But flexibility alone cannot eliminate the biological barrier between an artificial material and living neural tissue, which is why the researchers highlight a further step: biohybrid neural interfaces. These incorporate living biological components, including neural stem cells, neural progenitor cells and other neural cells, directly into the implanted device. The goal is twofold: improve integration with host tissue so the interface survives longer, and allow living tissue to participate in signal transmission itself. More advanced designs can even guide axon growth, creating new biological connections between neural tissue and the electronic device. In the most optimistic framing, biohybrid interfaces could extend beyond better electrodes to repairing damaged neural circuits, combining electronics, living cells and host tissue into a single functioning system.

None of this, the review stresses, makes ICMS a plug-and-play technology. Long-term performance depends on the stability of implanted electrodes, and stimulation effects vary between individuals and can drift over time within the same person. The authors therefore call for coordinated progress across electrode design, stimulation encoding, closed-loop calibration and safety evaluation, together with longer follow-up periods, cross-species validation, standardized safety assessments and reproducible behavioral and neural-network outcomes. Rather than replacing existing neuromodulation technologies, they argue, ICMS may ultimately serve as a complementary tool offering far finer control over local neural populations than surface stimulation or pharmacological approaches can achieve. Its long-term promise lies in combining several capabilities at once: delivering artificial sensory information, letting the brain learn entirely new information channels, reshaping dysfunctional circuits and integrating electronic devices more naturally with living neural tissue. As the review concludes, translating intracortical microstimulation from an experimental technique into durable brain-computer interface systems will require progress not in any single technology, but simultaneously across interface reliability, stimulation encoding, closed-loop control, safety and biohybrid integration. The work was supported by the National Key Research and Development Program of China, the Major Program of the National Natural Science Foundation of China and the National Natural Science Foundation of China.

Subject of Research: Intracortical microstimulation as a technique for evoking artificial perception and inducing plasticity in brain-computer interfaces

Article Title: Intracortical microstimulation in brain–computer interfaces: Evoking perception and plasticity

Article References: Intracortical microstimulation in brain–computer interfaces: Evoking perception and plasticity. (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: intracortical microstimulation, brain-computer interfaces, sensory feedback, phosphenes, neuroplasticity, closed-loop neuromodulation, biohybrid neural interfaces, somatosensory cortex, visual prosthesis, neural electrodes, Intracortical, microstimulation

The team designed a behavioral paradigm that allowed them to measure and even influence how participants made perceptual decisions across two distinct tasks. In the setup, animals and human participants had to judge stimuli along two different feature dimensions, and critically, they had to figure out on their own which of the two tasks was currently in effect. Because the task rule changed unpredictably, participants had to maintain and update an internal belief about which task was relevant, informed by whether previous trials had been rewarded. This design meant that the experimenters could track, trial by trial, how confident each participant was about the current task and how that confidence shaped the quality of their perceptual judgments.

The behavioral results were striking and consistent across species. Both humans and monkeys made less accurate perceptual decisions when the task was uncertain, and both were slower to respond under those conditions. An ideal observer analysis showed that this drop in perceptual accuracy was not a necessary consequence of switching between tasks; a perfectly rational agent could, in principle, change its task belief without sacrificing perceptual precision. Yet the biological participants consistently did. This mismatch between what is achievable and what brains actually do set the stage for the central question of the study: what, mechanistically, causes perception itself to degrade when the task is ambiguous?

To generate mechanistic hypotheses, the researchers turned to recurrent neural networks. They trained two types of networks on the same tasks the monkeys performed. One network, called the correct choice network, was trained simply to produce the best possible answers. The other, called the monkey choice network, was trained to replicate the actual, suboptimal choices of the monkeys, including their delays in switching tasks and their errors. Remarkably, only the network trained to mimic the monkeys reproduced the uncertainty-related drop in perceptual accuracy. The correct choice network switched tasks efficiently and maintained high perceptual performance throughout. This dissociation suggested that the behavioral cost was not an unavoidable property of the task architecture but instead reflected something specific about how biological brains manage uncertainty.

Peering inside the trained networks revealed the mechanism. Using distance covariance analysis, the team quantified how strongly the two stimulus features, the relevant and the irrelevant one, were represented in the network’s hidden layer activity during the interval between stimuli. In the monkey choice network, after unrewarded trials that lowered task certainty, the representation of the irrelevant feature became substantially stronger and the two feature representations lost their orthogonality, becoming entangled with one another. In the correct choice network, by contrast, the irrelevant feature was suppressed and the two feature axes remained nearly orthogonal even under high uncertainty. In other words, the network that replicated animal behavior literally mixed together the neural codes for features that should have been kept separate.

Armed with this hypothesis, the researchers turned to the brain itself. Recording from populations of neurons in the primary visual cortex, V1, of monkeys performing the task, they found the same signature of feature interference. Under high task certainty, V1 population activity primarily encoded the believed relevant feature, and information about the irrelevant feature decayed rapidly after stimulus offset. Under low task certainty, however, the irrelevant feature lingered in the population activity, and the neural representations of the two features became less separable. Decoders trained to read out feature values from V1 performed far better on the irrelevant feature during uncertain trials, mirroring the pattern seen in the monkey choice network and providing a direct neuronal correlate of the behavioral cost.

The team then pushed the analysis further, asking whether the animal’s own perceptual confidence could be decoded from V1 and used to predict behavior. The strength of the V1 representation of feature changes predicted how likely the monkey was to switch tasks after an unrewarded trial: when population activity indicated a strong, obvious change in the believed relevant feature, task switches were more frequent. Even among trials that were behaviorally identical, the decoded confidence from visual cortex carried information about the animal’s subsequent decision. This finding links a precise, measurable property of early sensory cortex to the higher-order process of deciding which task to perform, bridging perception and metacognition within a single neural circuit.

Crucially, the researchers went beyond correlation. Through behavioral experiments in humans, physiological recordings, and causal experiments using microstimulation in monkeys, they demonstrated that feature interference actively causes errors under uncertain conditions rather than merely accompanying them. Strengthening the representation of irrelevant features, whether through the natural dynamics of task uncertainty or through direct manipulation of the circuit, biased perceptual choices in predictable directions. Human participants showed idiosyncratic but systematic couplings between irrelevant and relevant feature judgments, and the monkeys exhibited similar directional interference across sessions, reinforcing the idea that the effect reflects a genuine representational mechanism rather than a generic lapse in attention or motivation.

The broader implications of the study reach into one of the oldest questions in psychology: why are cognitive capacities limited? The authors propose that capacity limitations, whether in multitasking, working memory, or attention, may stem fundamentally from interference between neural representations of different stimuli, tasks, or memories. When the brain cannot fully suppress one representation, it bleeds into another, and the resulting entanglement degrades both. This framework reframes the classic switch costs documented in decades of task-switching research as a consequence of representational geometry: under certainty, the brain maintains orthogonal codes for different features and tasks, but uncertainty erodes that orthogonality, allowing signals to contaminate one another.

The study also showcases a methodological blueprint that is likely to influence the field. By training one network to be optimal and another to match animal behavior, and then comparing their internal representations, the researchers could generate a falsifiable mechanistic hypothesis and test it with both observational and causal tools in biological brains. The monkey behavioral, electrophysiology, and microstimulation data, together with the human psychophysics data, have been made openly available, and the code for network training and analyses has been released to the research community. As artificial neural networks increasingly serve as models of cognition, this work demonstrates both the power and the subtlety of that approach: networks can reveal not only what brains do, but, when trained on behavior rather than on correctness, why brains fall short of their own impressive potential.

Subject of Research: Neuronal mechanisms of task uncertainty and feature interference in perceptual decision-making

Article Title: Feature interference underlies a neuronal basis for the behavioral cost of task uncertainty

Article References: Xue, C., Markman, S. K., Chen, R., Kramer, L. E., & Cohen, M. R. (2026). Feature interference underlies a neuronal basis for the behavioral cost of task uncertainty. Nature Neuroscience. https://doi.org/10.1038/s41593-026-02430-w

Image Credits: AI Generated

DOI: 10.1038/s41593-026-02430-w

Keywords: task uncertainty, feature interference, decision-making, visual cortex, neural networks, perception, task switching, neuroscience, Nature Neuroscience, cognitive capacity, microstimulation, psychophysics