Summary: Researchers have developed a two-photon imaging platform called FlatMux that records electrical voltage signals from nearly 200 neurons simultaneously in living brain tissue, with millisecond precision and at depths up to 500 micrometers.
FlatMux rethinks optical light delivery to address the competing challenges of laser power, tissue heating, and detection sensitivity. Using an engineered optical cavity to split a laser into precisely timed, non-overlapping pulses rather than relying on mechanical raster scanning, the system maximizes fluorescence photon efficiency. This enables frame rates up to 2,000 frames per second while keeping excitation energy low enough to avoid excessive heating and photodamage.
Key Facts
- Unprecedented speed and scale: FlatMux captures direct electrical activity from nearly 200 neurons at once, across single or multiple cortical layers, with frame rates up to 2,000 Hz.
- Deep tissue penetration: The platform detects millisecond voltage signals at depths up to 500 micrometers in scattering cortical tissue of awake, head-fixed mice.
- Efficient optical cavity engineering: Rather than mechanically sweeping a laser beam, FlatMux splits laser pulses into precisely timed, non-overlapping points, sampling at an effective rate of 150 million points per second to maximize photon return per excitation photon and reduce heating.
- Sub-threshold signal detection: In high-sensitivity mode, FlatMux resolves small voltage fluctuations that precede action potentials, enabling optical mapping of synaptically connected microcircuits.
- Cell-type and location information: Direct voltage imaging via fluorescent genetically encoded voltage indicators (GEVIs) provides simultaneous information on physical cell location, cell type, and real-time function without inserting electrodes.
Source: Rockefeller University
Overview: The brain is a densely interconnected network of neurons that compute in parallel. Observing those computations in real time is critical to understanding how sensation is processed and how behavior is guided. Electrically, these processes occur on millisecond timescales and are difficult to measure in living tissue because the signals are fast, faint, and easily obscured by noise or photodamage.
A study published in Nature Methods introduces FlatMux, a two-photon imaging platform designed to record voltage signals from large neuronal populations with high temporal precision. By minimizing the excitation energy required to detect each neuron and by adopting a scalable optical scanning strategy, the system records nearly 200 neurons across one or more planes, reaches imaging depths of up to 500 µm, and operates at frame rates up to 2,000 Hz.
FlatMux combines hardware and software: an optical design that maximizes fluorescent photons returned per excitation photon, and computational methods that separate true neural signals from background noise. Together, they enable recording of fast voltage dynamics—including both spiking and sub-threshold events—across distributed networks in awake, behaving animals.
Unlike calcium imaging, which reports neural activity indirectly and with slow kinetics, voltage imaging with GEVIs reports membrane potential directly on millisecond timescales. However, GEVIs are fainter and require two-photon excitation to image deep tissue, creating a trade-off between signal quality and tissue safety. Traditional two-photon systems tend to oversample areas during mechanical scanning, wasting energy and increasing heating. FlatMux addresses these challenges by delivering light only where and when it is needed.
How FlatMux works: The platform uses an optical cavity to split a pulsed laser into many precisely timed pulses that are routed to distinct, non-overlapping points in the tissue. This lateral-temporal multiplexing avoids redundant illumination and concentrates excitation where fluorescence will be measured, improving photon efficiency and reducing heating. The design is flexible: modes include a large field-of-view setting, a 2-kHz high-speed mode, a deep-tissue mode for imaging up to 500 µm, a dual-plane mode for simultaneous recordings in multiple cortical layers, and a high-SNR mode optimized for detecting sub-threshold voltage fluctuations while minimizing pixel crosstalk and photobleaching.
In experimental tests, the research team used FlatMux to image awake mice running on a treadmill while delivering whisker stimulation. The system recorded faint voltage responses during sensory stimulation and spontaneous behavior, demonstrating the capacity to track rapid electrical events and to observe how signals propagate through layered cortical circuits in real time.
FlatMux recorded almost 200 neurons simultaneously, reached depths of 500 µm, captured activity at up to 2,000 frames per second, and could image two cortical layers at once. In its high-sensitivity mode it resolved sub-threshold signals that reveal synaptic coupling between neurons, suggesting the possibility of inferring functional connectivity optically and at scale.
By enabling large-scale, millisecond-resolution voltage imaging, FlatMux may shift neuroscience from isolated snapshots of single cells to continuous recordings of distributed neural computation. This capability can reveal how activity flows through circuits during perception, decision-making, and behavior, and may offer a new way to classify cell types by their electrical signatures.
“For a long time the field focused on response properties of single neurons,” says Alipasha Vaziri, head of the Laboratory of Neurotechnology and Biophysics at Rockefeller University. “FlatMux makes it possible to investigate how brain functions arise from distributed information processing across interconnected networks.”
Key Questions Answered:
A: Calcium signals are indirect and slow, unfolding over hundreds of milliseconds, while voltage signals occur in 1–2 milliseconds. Voltage imaging therefore requires much faster recording speeds and must detect fainter fluorescent signals from GEVIs, which is challenging in deep, scattering tissue without causing damage.
A: Traditional two-photon systems sweep a laser across tissue and often oversample regions, depositing unnecessary energy. FlatMux uses an optical cavity to deliver discrete, non-overlapping pulses only where needed, increasing photon efficiency per excitation and reducing excess heating.
A: Sub-threshold responses are small membrane potential fluctuations that occur before a neuron fires. Detecting them reveals how inputs are integrated across synaptically connected neurons and provides insight into functional circuit wiring beyond spike timing alone.
Editorial Notes:
- This article was edited by a Neuroscience News editor.
- The journal paper was reviewed in full.
- Additional context was added by our editorial staff.
About this neurotech research news
Author: Katherine Fenz
Source: Rockefeller University
Contact: Katherine Fenz – Rockefeller University
Image: The image is credited to Neuroscience News
Original Research: Open access. “A versatile platform for two-photon neuronal population voltage imaging across cortical depths” by Jingkun Guo, Kevin Barber, M. Agustina Frechou, Sihao Lu, Jeff Demas, David Chen, Shuyuan Yang, Alex James McDonald, Michelle Ann Land, François St-Pierre & Alipasha Vaziri. Nature Methods. DOI: 10.1038/s41592-026-03158-y
Abstract
A versatile platform for two-photon neuronal population voltage imaging across cortical depths
Genetically encoded voltage indicators allow direct, high spatiotemporal resolution measurements of neuronal membrane potential in genetically defined cell populations. Their fast dynamics, low signal-to-noise ratio, and photobleaching have limited wider adoption and made scaling to large neuronal populations difficult, especially when combined with suboptimal optical acquisition schemes.
FlatMux is a flexible lateral-temporal multiplexing platform that provides a scalable, spatiotemporally and energetically efficient two-photon imaging scheme. The system is reconfigurable to meet different recording needs: it offers a large field-of-view mode, a 2-kHz high-speed mode, a deep-tissue mode for cortical recordings down to 500 µm, a dual-plane mode for simultaneous recordings in multiple cortical layers, and a high-SNR mode for resolving sub-threshold and high-SNR spiking activity while minimizing pixel crosstalk and bleaching.
By maximizing photon efficiency and enabling recording of single-spike and single-trial activity across large neuronal populations, FlatMux addresses key challenges in multiphoton voltage imaging and opens new opportunities for studying complex brain functions in vivo.