Precision Control: Advances in Holographic Stimulation for Two-Photon Optogenetics

Earlier this year, Filipe Machado Ferreira aptly likened the brain to "the harmonious coordination of an orchestra in full performance," a metaphor that vividly captures the intricate symphony of neural activity. Just as an orchestra employs polyrhythm—where multiple rhythms, each with distinct beats, weave together into a cohesive soundscape—the brain’s networks generate complex, overlapping spatiotemporal patterns. These patterns emerge from the unique interplay of neural circuit structures, shaped by intrinsic morphology, physiological properties, and the adaptive transformations occurring over a lifetime of experiences and stimuli. Each "note" of neural firing builds upon others, enabling the finely tuned coordination required for processing information seamlessly each day.

Understanding the interactions between neurons within a neural circuit is central to the appeal of optogenetics. In vivo calcium imaging provides an indirect yet insightful readout of neural activity¹, while channel rhodopsin-based photoactivation enables precise manipulation of neuronal firing^2,3 (Figure 1a). Through this approach, researchers can activate or inhibit specific neurons by exploiting cellular mechanisms and ion gradients, allowing them to observe the effects within the context of an interacting circuit. By combining in vivo calcium imaging with targeted photo stimulation, scientists can perturb individual neurons within a circuit under physiological conditions, uncovering the fundamental mechanisms and computational processes that support local circuit dynamics. This approach provides an unprecedented window into how neural networks coordinate the large-scale transfer of information across space and time, illuminating the sophisticated patterns that drive coherent brain activity.

Returning to the concept of brain networks as complex, overlapping spatiotemporal patterns—each with its own rhythms and distinct beats—an unbiased perturbation of multiple circuits across a large area is unlikely to achieve the physiological specificity needed to distinguish intricate differences in circuit-specific neural activity. Although stimulation power and frequency can be carefully controlled, all neurons expressing channelrhodopsins are subject to the same perturbation dynamics, potentially overlooking the unique, rhythmically nuanced interactions within each circuit. This lack of specificity underscores the challenge of precisely tuning perturbations to match the natural complexity of individual networks.

Enter the application of spatial light modulators (SLMs) in microscopy—a ground-breaking system capable of phase-modulating a laser beam to generate any spatiotemporal light pattern, enabling simultaneous illumination of multiple selected regions of interest^5,6. This technology allows a single 2-photon beam to be split into spatial patterns, targeting numerous channelrhodopsin-expressing cells within a field of view, with each coordinate receiving light with 2-photon specificity⁷.

An SLM, such as Meadowlark’s 1024x1024 model, operates through an array of rotatable liquid crystal beads that adjust their orientation and angle based on applied voltage, generating precise spatial light patterns (Figure 2). This device can simultaneously target up to 375 distinct locations within a ~200 µm column of tissue. When combined with galvanometer mirror movements, the effective field of view extends to ~600 µm through sequential stimulation. The system supports 2-photon calcium imaging alongside spatially patterned photostimulation, with an SLM refresh rate of 1.7 milliseconds and galvanometer repositioning within 3 milliseconds. Together, these capabilities allow the activation of thousands of neurons within a 10-millisecond window, delivering unparalleled precision and scalability in all-optical neural circuit interrogation.

Once a cell type of interest and an experimental paradigm have been established, it becomes crucial to understand the intrinsic functional properties of the cell type to design perturbation strategies that align with biologically relevant firing frequencies and dynamics. Two-photon stimulation beams are rarely exposed to tissue without some form of pulsing. For example, most SLM-enabled photostimulation setups inherently pulse at 266 femtoseconds, but further adjustments to parameters such as frequency of exposure, duration of exposure, and repetition count are often required to achieve a physiologically relevant stimulation profile.

These optimizations are critical to ensure that perturbations effectively mimic natural activity patterns and provide meaningful insights into the function and dynamics of the targeted neural circuits. For instance, chandelier cells, which are fast-spiking neurons, require trains of rapid action potentials at a rate of 40–200 Hz. In contrast, Purkinje cells fire at relatively slower rates of 20–100 Hz. These differences necessitate fundamentally distinct stimulation strategies, including variations in power output, duration of exposure, and repetition count of the beam activity. Tailoring these parameters to the specific firing dynamics of each cell type is essential for accurate and biologically relevant experimental outcomes.

Optimizations must extend beyond the firing properties of specific cell types to include their functional relevance. If your cell type exhibits temporally unique firing patterns—whether across a behavioral paradigm or during spontaneous activity—these patterns can be leveraged to create a functional mapping of cells.

In the cerebellum, for instance, Purkinje cells are organized into spatial structures known as microzones, which receive feedforward input from the same source neuron and relay coordinated output to the deep cerebellar nuclei. Functional metrics, such as calcium imaging, can be employed to identify these microzones. Research indicates that Purkinje cell dendrites within a single microzone exhibit synchronized activity within the same temporal window, a characteristic that differentiates them from adjacent microzones¹¹.

This functional organization highlights the necessity of tailoring stimulation strategies not only to the intrinsic firing properties of individual cells but also to the broader spatial and temporal dynamics of the circuits they belong to. By understanding the temporally unique roles of individual microzones relative to their neighbors, experimenters can group Purkinje cells into functionally relevant stimulation targets. This approach helps avoid the critical mistake of stimulating two functionally distinct microzones with a single perturbation, thereby preserving the integrity of circuit-specific investigations (Figure 4).

Once you have developed a strong understanding of the neural dynamics in your system and the role your cell type of interest plays within these spatiotemporal dynamics, an exciting opportunity arises: performing real-time closed-loop control of your circuit’s activity¹⁴. By understanding the temporal evolution of neural activity and its relationship to specific biological or behavioral outcomes, you can design precise perturbations to influence these outcomes.

What types of perturbations could be applied, and at what time intervals, to drive the organism to choose or act differently? How consistently can these manipulations produce the desired changes? These questions lie at the frontier of neuroscience, allowing researchers to probe causality and explore the functional flexibility of neural circuits in real time.

The outcomes that feel most significant and relevant to me in this research lie in its potential applications, such as neuroprosthetics, brain-computer interfaces, and alleviating the effects of epileptic seizures through real-time prediction and closed-loop intervention. While these align closely with my personal interests, I am confident that there are countless other applications yet to be explored, waiting to be unlocked by the discovery of information from this technology.

I hope that through reading this, you have gained a clear starting point for conducting all-optical interrogations of neural circuits, including the application of SLM-enabled systems for single-cell-specific probing within larger, more complex neural networks. Moreover, I hope this has provided insight into the fundamental biological, behavioral, and computationally relevant outcomes you can explore using this powerful experimental protocol. With this approach at your disposal, the potential to unravel the intricate dynamics of neural circuits and their contributions to system-level functions is within reach.