The device, which the authors call an ionic-electronic event-driven sensor, or IEES, is built from a 130-nanometre-thick bulk heterojunction of two materials: PgBDT-T, a pi-conjugated polymer that belongs to the family of organic mixed ionic-electronic conductors, and Y6, a well-known non-fullerene acceptor. The blend is coated onto a gold working electrode and immersed in an electrolyte that contacts a counter electrode. What makes the architecture remarkable is its simplicity. Conventional photosensors that need both fast temporal response and adaptive gain typically require multiple transistors per pixel, with separate components handling detection, spike generation and modulation. The IEES achieves all three functions within one photoactive film, because the two blended materials divide the labour naturally.

The operating mechanism hinges on a stark mobility asymmetry between the two phases. Space-charge-limited current measurements show that electrons move through Y6 with a mobility of 7.0 x 10^-4 square centimetres per volt-second, more than two orders of magnitude faster than holes move through PgBDT-T, whose hole mobility is just 5.6 x 10^-6. When light strikes the blend, excitons generated at the heterojunction interface dissociate almost completely, as confirmed by near-total photoluminescence quenching in the optimised 1:1 blend. Electrons race through the Y6 network to the gold electrode, producing a sharp positive photocurrent spike. Meanwhile, the holes left behind in the sluggish PgBDT-T phase accumulate as positive charge.

Here is where the ionic chemistry takes over. Because PgBDT-T is ion-accessible, mobile anions from the electrolyte penetrate the polymer and compensate the accumulating positive charge. This ion-compensated charging weakens the local electric field that extracts photocarriers, promotes recombination and suppresses any further photocurrent. The result is a self-terminating spike: a fast excitatory onset followed by intrinsic inhibitory feedback, functionally analogous to the way inhibitory interneurons in the retina rein in ganglion-cell firing after an initial burst. When the light switches off, the stored charge is released, producing a characteristic negative spike. The device then returns to an electrically quiescent state with no detectable photocurrent, which means near-zero standby power.

The team marshalled an impressive battery of evidence to pin down this mechanism. In operando ultraviolet-visible spectroscopy revealed progressive bleaching of the PgBDT-T absorption near 540 nanometres under increasing illumination, accompanied by a weak sub-gap absorption band, the spectroscopic fingerprint of p-type charging. Kelvin probe force microscopy showed surface-potential shifts in the same direction whether the films were illuminated or electrochemically oxidised. Quartz crystal microbalance measurements documented swelling-associated anion uptake in PgBDT-T while Y6 remained essentially ion-inert. Control devices built from the non-OMIEC blend PTQ10:Y6, tested under identical geometry, electrolyte and illumination, showed only slow monotonic photocurrent decay with no spike at all, confirming that ion-accessible mixed conduction is the essential ingredient. Drift-diffusion simulations reinforced the picture: as the effective charge density rose from 10^16 to 10^18 per cubic centimetre, the simulated short-circuit current collapsed from roughly 95 percent to 17 percent of its neutral value, because bimolecular recombination and transport bottlenecks throttled extraction.

The sensor’s performance figures are striking in their own right. The photocurrent spike amplitude follows a power law in the illumination change, with exponents of about 0.28 for turn-on events and 0.05 for turn-off events across a range spanning 1 to 90 milliwatts per square centimetre. The minimum detectable change is 0.055 milliwatts per square centimetre, roughly 60 lux, comparable to the lower comfort limit for human reading, and the dynamic range exceeds 75 decibels. The device operated with less than 5 percent deviation over 20,000 seconds of continuous cyclic illumination at zero bias in ambient laboratory conditions, suggesting that the ionic feedback is genuinely reversible rather than degrading. Broadband operation was confirmed at wavelengths of 455, 565, 660 and 850 nanometres, and the response tracks the rate of illumination change, with fast ramps producing narrow, high-amplitude spikes that saturate above about 100 milliwatts per square centimetre per second.

The most consequential feature, however, is voltage-programmable amplitude control. Applying a small positive bias to the working electrode strengthens the ionic feedback pathway through bias-assisted anion compensation, and the amplitude coefficient of the spike response falls exponentially with voltage, with decay constants of 7.2 per volt for turn-on events and 11.1 per volt for turn-off events, while the power-law exponents remain largely untouched. In other words, bias rescales the excitation-inhibition balance without distorting the underlying contrast sensitivity. Background illumination provides a complementary passive adaptation, attenuating spikes exponentially with increasing background level. The team demonstrated the practical payoff with a 4 x 4 pixel array on which dynamic letter patterns were projected: applying 0.2 volts to brightly lit pixels and zero volts to dim ones normalised the responses into a comparable amplitude range, an active gain equalisation performed entirely within the sensor.

Scaled to a 128 x 128 pixel simulation of a video containing three moving participants, the same principle produced something resembling visual attention. Under a uniform 0.1-volt bias, event outputs simply reflected local motion. But when the researchers assigned zero volts to one participant, 0.2 volts to another and 0.1 volts elsewhere, the response from the first participant was enhanced nearly tenfold while the second was strongly suppressed, with the third essentially unchanged. Because the convolutional kernel weights of a neural network can likewise be encoded as pixel-specific bias voltages, the array performs analogue weighting physically, before any data leave the sensor. The authors call this hybrid amplitude-temporal coding: each illumination change generates one analogue event whose magnitude carries both the local contrast change and the programmed weight.

The system-level demonstration is where the efficiency argument becomes concrete. Feeding device-informed event streams from the Weizmann human-action dataset into a lightweight long short-term memory classifier, the IEES-based pipeline achieved 100 percent classification accuracy on four action classes, compared with 94.2 percent for a frame-based network and 75.8 percent for an event-camera-based network using the same downstream architecture. More telling is the energy accounting. Under matched benchmarking conditions across the full 149-video set, the IEES front end consumed 0.064 millijoules, including event generation and in-sensor analogue convolution. The event-camera pipeline required 0.69 joules and the frame-based pipeline 2.80 joules, while an RGB rate-coding baseline demanded 15.70 joules. That is a difference of more than four orders of magnitude, arising because amplitude coding preserves motion-relevant contrast in a single sparse event rather than spending spike counts to represent magnitude.

The platform also shows encouraging generality. Event-driven responses persisted in sodium, calcium and potassium carbonate electrolytes, in solid-state devices using sodium chloride hydrogel, across active areas from 0.24 to 1.0 square centimetre, and in microfabricated pixel arrays with consistent polarity and timing. Alternative OMIEC donors including PgBDT-Se, PTQ-6bO2, HOMO-gDPP, PgDPP-TT and Pg2TE-TT-i all produced clear spikes when blended with Y6, while the closely related but ion-inaccessible P3HT:Y6 blend showed only monotonic decay, underscoring ionic accessibility as the decisive design rule. Much work remains before such electrolyte-interfaced devices can rival mature silicon imagers in integration density and durability, and the amplitude variation across manually dispensed electrolyte droplets hints at the manufacturing challenges ahead. But the conceptual advance is clear and potentially far-reaching: by recruiting mobile ions as an intrinsic inhibitory mechanism, organic mixed conductors offer a materials-level route to vision hardware that senses, adapts and prioritises in one place, the way biology has always done it.

Subject of Research: Organic mixed ionic-electronic conductor photosensors with ionic-mediated inhibitory feedback for event-driven neuromorphic vision

Article Title: Amplitude-controllable event-driven organic photosensors based on ionic-mediated inhibition

Article References: Zhao, C., Su, X., Chen, X., Zhang, S., Yuan, Z., Lin, N., Shen, J., Tao, Y., Wang, S., Wang, B., Li, T., Li, X.-H., Fan, Q., Wang, Z., & Ma, W. (2026). Amplitude-controllable event-driven organic photosensors based on ionic-mediated inhibition. Nature Materials. https://doi.org/10.1038/s41563-026-02747-8

Image Credits: AI Generated

DOI: 10.1038/s41563-026-02747-8

Keywords: organic mixed ionic-electronic conductors, neuromorphic vision, event-driven sensing, bulk heterojunction, ionic-electronic coupling, temporal-contrast detection, PgBDT-T, Y6, spike amplitude modulation, in-sensor computing, retina-inspired sensors, low-power photodetection

Conventional frame cameras, the dominant technology in everything from smartphones to industrial inspection systems, sample the visual field at fixed intervals. Each frame is a dense snapshot: every pixel reports a brightness value, regardless of whether anything in the scene has changed. This uniform sampling makes frames easy to store, compress, and process with the deep convolutional networks that power contemporary computer vision. But it also imposes costs. Between exposures, motion is invisible; during exposure, fast motion produces blur; and the fixed frame rate forces systems to allocate identical computational effort to static scenes and dynamic ones, wasting energy and latency budget where nothing of interest is happening.

Event cameras, also known as dynamic vision sensors or neuromorphic cameras, invert this logic. Rather than reporting absolute brightness at fixed intervals, each pixel independently and asynchronously emits an event whenever the local logarithmic brightness changes by a threshold amount. The result is a sparse, temporally precise stream: a spinning fan or a flickering LED generates dense event activity, while a motionless wall generates almost none. Event sensors offer microsecond-scale temporal resolution, very high dynamic range, and low power consumption, and they suppress redundancy by construction. Yet they have their own weaknesses. Absolute appearance information is absent, textures that produce no brightness change generate no signal, and the spiking nature of the data resists the standard toolkits of frame-based deep learning.

The central contribution of the new study is to treat these complementary behaviors not as anecdotal engineering observations but as a measurable modality advantage. The authors analyze the conditions under which each modality contains information the other cannot supply, and they show that the advantage of hybrid data emerges from the statistical structure of natural scenes: real environments mix static structure, which frames capture efficiently, with sparse dynamic events, which event streams capture with far greater temporal fidelity. When both representations are available for the same scene, the combined data constrains the underlying visual state more tightly than either stream alone, and the study formalizes when and by how much.

This framing has immediate practical consequences. In many deployed hybrid systems, the event stream is treated as a helper signal, converted into synthetic frames or used to deblur and interpolate the primary frame data. The new analysis suggests that such designs may systematically underuse the event modality. If the fundamental advantage of events lies in their asynchronous, change-driven sampling, then forcing them into a frame-like representation discards precisely the property that makes them valuable. The authors argue for architectures that preserve the native temporal structure of event data and fuse it with frames at the level of information content rather than at the level of pixel grids.

The study also addresses the inverse question: what do frames contribute that events cannot? Dense appearance, texture, color, and absolute illumination are all naturally carried by frames and are difficult or impossible to recover from events alone. In low-texture scenes, slow-motion regimes, or conditions where brightness changes fall below the sensor’s event threshold, the frame modality holds the decisive information. A rigorous account of modality advantage therefore predicts not a universal winner but a regime-dependent division of labor, with the balance shifting as scene dynamics, lighting conditions, and motion speeds change. This regime-dependent view gives engineers a principled basis for deciding, task by task, how much weight to assign each stream.

Applications stand to gain across several domains. In autonomous driving and robotics, where latency and power budgets are unforgiving, hybrid sensing promises reliable perception across the full range of operating conditions: frames anchor recognition when the scene is static or slowly changing, while events carry the microsecond-level motion cues needed for fast obstacle avoidance, high-speed tracking, and operation in challenging lighting such as tunnel exits or night driving. In industrial monitoring, the same logic applies to vibration analysis, high-speed defect detection, and inspection of machinery whose moving parts would blur in conventional video. In scientific imaging, hybrid acquisition can capture transient phenomena with event precision while retaining frame-based photometric context.

The analysis also speaks to a growing body of work on learning from hybrid data. Training models that consume both frames and events raises questions of representation alignment, since the two modalities differ in geometry, sampling density, and temporal structure. The study’s account of fundamental modality advantage offers a criterion for evaluating fusion strategies: a good hybrid model should extract information that is provably unavailable to single-modality models, rather than merely averaging redundant estimates. The authors discuss how this criterion can guide dataset design and benchmark construction, encouraging evaluations that specifically probe the regimes where hybrid sensing should outperform, such as high-speed motion, extreme dynamic range, and low-power operation.

Challenges remain before the framework’s implications fully reach practice. Event cameras are still less widespread and less standardized than frame sensors, hybrid hardware remains costly, and the algorithms best suited to asynchronous data continue to evolve rapidly, drawing on tools from spiking neural networks, graph-based processing, and temporal deep learning. Benchmark datasets that simultaneously record both modalities with precise time alignment are still relatively scarce, which complicates the kind of regime-mapped evaluation the study advocates. Nevertheless, the work provides a conceptual scaffold for the field: rather than treating frame-event fusion as a bag of engineering tricks, it offers a theory of when and why the combination helps, grounded in the information structure of the visual world.

As neuromorphic sensors mature and hybrid camera systems move from laboratories into commercial products, the question of modality advantage will only grow in importance. The new study positions itself as a step toward answering that question rigorously, giving researchers and engineers a shared vocabulary for reasoning about what each modality contributes, where the combination is indispensable, and how future vision systems should be built to exploit both. If its central claim holds across tasks and hardware platforms, the era of choosing between frames and events may give way to something more interesting: systems designed from the outset to harvest the fundamental advantages of both.

Subject of Research: Fundamental modality advantage in hybrid frame and event-based visual data

Article Title: Understanding and exploiting fundamental modality advantage in frame and event hybrid visual data

Article References: Wang, S., Zheng, H., Xia, H., Wang, Z., Qi, X., Han, X., Wang, X., Wu, J., & Deng, L. (2026). Understanding and exploiting fundamental modality advantage in frame and event hybrid visual data. Communications Engineering. https://doi.org/10.1038/s44172-026-00778-2

Image Credits: AI Generated

DOI: 10.1038/s44172-026-00778-2

Keywords: event cameras, neuromorphic vision, frame-based imaging, hybrid visual data, modality advantage, machine vision, dynamic vision sensors, sensor fusion, computer vision, low-power sensing, autonomous robotics, visual information processing