Voltage imaging holds great potential for biomedical research by enabling noninvasive recording of the electrical activity of excitable cells such as neurons or cardiomyocytes. Camera-based detection can record from hundreds of cells in parallel, but imaging entire volumes is limited by the need to focus through the sample at high speeds. Remote focusing techniques can remedy this drawback, but have so far been either too slow or light-inefficient. Here, we introduce flipped image remote focusing, a remote focusing method that doubles the light efficiency compared to conventional beamsplitter-based techniques and enables high-speed volumetric voltage imaging at 500 volumes/s. We show the potential of our approach by combining it with light sheet imaging in the zebrafish spinal cord to record from >100 spontaneously active neurons in parallel.

Voltage imaging measures neuronal activity directly and holds promise for understanding information processing within individual neurons and across populations. However, imaging voltage over large neuronal populations has been challenging owing to the simultaneous requirements of high imaging speed and signal-to-noise ratio, large volume coverage and low photobleaching rate. Here, to overcome this challenge, we developed a confocal light-field microscope that surpassed the traditional limits in speed and noise performance by incorporating a speed-enhanced camera, a fast and robust scanning mechanism, laser-speckle-noise elimination and optimized light efficiency. With this method, we achieved simultaneous recording from more than 300 spiking neurons within an 800-µm-diameter and 180-µm-thick volume in the mouse cortex, for more than 20 min. By integrating the spatial and voltage activity profiles, we have mapped three-dimensional neural coordination patterns in awake mouse brains. Our method is robust for routine application in volumetric voltage imaging.

In calcium imaging studies, Ca transients are commonly interpreted as neuronal action potentials (APs). However, our findings demonstrate that robust optical Ca transients primarily stem from complex "AP-Plateaus", while simple APs lacking underlying depolarization envelopes produce much weaker photonic signatures. Under challenging in vivo conditions, these "AP-Plateaus" are likely to surpass noise levels, thus dominating the Ca recordings. In spontaneously active neuronal culture, optical Ca transients (OGB1-AM, GCaMP6f) exhibited approximately tenfold greater amplitude and twofold longer half-width compared to optical voltage transients (ArcLightD). The amplitude of the ArcLightD signal exhibited a strong correlation with the duration of the underlying membrane depolarization, and a weaker correlation with the presence of a fast sodium AP. Specifically, ArcLightD exhibited robust responsiveness to the slow "foot" but not the fast "trunk" of the neuronal AP. Particularly potent stimulators of optical signals in both Ca and voltage imaging modalities were APs combined with plateau potentials (AP-Plateaus), resembling dendritic Ca spikes or "UP states" in pyramidal neurons. Interestingly, even the spikeless plateaus (amplitude > 10 mV, duration > 200 ms) could generate conspicuous Ca optical signals in neurons. Therefore, in certain circumstances, Ca transients should not be interpreted solely as indicators of neuronal AP firing.

Sensory stimulations at 40 Hz gamma (but not any other frequency), have shown promise in reversing Alzheimer's disease (AD)-related pathologies. What distinguishes 40 Hz? We hypothesized that stimuli at 40 Hz might summate more efficiently (temporal summation) or propagate more efficiently between cortical layers (vertically), or along cortical laminas (horizontally), compared to inputs at 20 or 83 Hz. To investigate these hypotheses, we used brain slices from AD mouse model animals (5xFAD). Extracellular (synaptic) stimuli were delivered in cortical layer 4 (L4). Leveraging a fluorescent voltage indicator (VSFP) expressed in cortical pyramidal neurons, we simultaneously monitored evoked cortical depolarizations at multiple sites, at 1 kHz sampling frequency. Experimental groups (AD-Female, CTRL-Female, AD-Male, and CTRL-Male) were tested at three stimulation frequencies (20, 40, and 83 Hz). Despite our initial hypothesis, two parameters-temporal summation of voltage waveforms and the strength of propagation through the cortical neuropil-did not reveal any distinct advantage of 40 Hz stimulation. Significant physiological differences between AD and Control mice were found at all stimulation frequencies tested, while the 40 Hz stimulation frequency was not remarkable.

Voltage imaging is an important complement to traditional methods for probing cellular physiology, such as electrode-based patch clamp techniques. Unlike the related Ca imaging, voltage imaging provides a direct visualization of bioelectricity changes. We have been exploring the use of sulfonated silicon rhodamine dyes (Berkeley Red Sensor of Transmembrane potential, BeRST) for voltage imaging. In this study, we explore the effect of converting BeRST to diEt BeRST, by replacing the dimethyl aniline of BeRST with a diethyl aniline group. The new dye, diEt BeRST, has a voltage sensitivity of 40% Δ/ per 100 mV, a 33% increase compared to the original BeRST dye, which has a sensitivity of 30% Δ/ per 100 mV. In neurons, the cellular brightness of diEt BeRST is about 20% as bright as that of BeRST, which may be due to the lower solubility of diEt BeRST (300 μM) compared to that of BeRST (800 μM). Despite this lower cellular brightness, diEt BeRST is able to record spontaneous and evoked action potentials from multiple neurons simultaneously and in single trials. Far-red excitation and emission profiles enable diEt BeRST to be used alongside existing fluorescent indicators of cellular physiology, like Ca-sensitive Oregon Green BAPTA. In hippocampal neurons, simultaneous voltage and Ca imaging reveals neuronal spiking patterns and frequencies that cannot be resolved with traditional Ca imaging methods. This study represents a first step toward describing the structural features that define voltage sensitivity and brightness in silicon rhodamine-based BeRST indicators.