Three-dimensional multi-site random access photostimulation (3D-MAP)

Yi Xue, Laura Waller, Hillel Adesnik, Nicolas C. PégardeLife 2022 · DOI

TL;DR

This paper introduces 3D-MAP, a novel one-photon optogenetic technique that overcomes the limitations of existing methods by enabling high-resolution, high-speed, and cost-effective 3D photostimulation of large neural ensembles. It achieves this by dynamically modulating light in both spatial and angular domains, paving the way for more comprehensive brain circuit interrogation.

What Problem Does This Paper Address?

Understanding brain function and disease relies heavily on precisely controlling and observing neural activity. Optogenetics, a technique that uses light to control genetically modified neurons, is a powerful tool for this. However, current methods face significant trade-offs:

  • Two-photon holographic optogenetics: Offers high spatial resolution and depth penetration but requires very high laser power, limiting the number of neurons that can be stimulated simultaneously. It's also extremely expensive and complex, hindering widespread adoption.
  • One-photon techniques: Generally use lower power and are more cost-effective but struggle with achieving both high spatial resolution and large 3D volumes simultaneously. Existing one-photon methods include:
  • Scanning-based: Stimulates neurons sequentially, not suitable for simultaneous ensemble activation.
  • 2D projection (DMD): Lacks axial resolution and can cause unwanted stimulation above/below the focal plane.
  • Computer-generated holography (CGH): Limited by SLM refresh rates, complex computation for 3D patterns, and spatial cross-talk.

This leaves a gap for a technology that can achieve high-throughput, all-optical interrogation of large, 3D-distributed neural ensembles with both speed and precision, at a reasonable cost.

Methods

The core innovation is 3D-MAP (Three-Dimensional Multi-Site Random Access Photostimulation), a one-photon light sculpting technique that combines spatial and angular light modulation.

  • Key Components:
    • Digital Micromirror Device (DMD): Modulates light intensity in the spatial domain (x, y).
  • Scanning Mirrors: Control the angles of illumination (kx, ky).
  • How it Works: The system synchronizes the DMD, which projects amplitude masks, with the scanning mirrors, which sweep through specific angles. This allows for the dynamic generation of 3D light patterns. The "degrees of freedom" (DoF) are significantly increased by multiplying the DoF of the DMD and scanning mirrors, unlike previous methods that summed them.
  • Advantages:
    • High DoF: Enables precise control over a large 3D volume.
  • Reduced Cross-talk: By prioritizing illumination angles that avoid non-targeted areas.
  • High Throughput & Speed: DMD refresh rates (multi-kHz) allow for volumetric pattern refresh rates exceeding neural circuit bandwidth, significantly faster than CGH.
  • Cost-Effective: Significantly cheaper than two-photon systems.
  • Lower Power: Uses orders of magnitude less laser power, reducing brain heating concerns.
  • Computational Framework: A gradient descent algorithm (FISTA) is used for reconstructing synaptic connectivity maps from multi-site stimulation data, enabling real-time analysis.

Datasets

The paper does not discuss datasets used for training machine learning models. The experiments utilize in vitro brain slices and in vivo mouse brains as biological samples for photostimulation and recording.

Key Insights and Ideas

  • 4D Light Field Control: 3D-MAP is the first system to use a 4D light field (x, y, kx, ky) for in vivo optogenetic photostimulation, offering unprecedented control over light delivery.
  • DoF Multiplication: The key to 3D-MAP's power lies in multiplying the degrees of freedom of the spatial modulator (DMD) and angular modulators (scanning mirrors), leading to a vastly larger controllable space compared to previous additive approaches.
  • Scalability and Cost-Effectiveness: The combination of high performance (speed, resolution, volume) with significantly lower cost makes 3D-MAP accessible to a broader range of neuroscience labs.
  • All-Optical Interrogation: The integration of photostimulation with fluorescence imaging allows for simultaneous manipulation and recording of neural activity in vivo, enabling high-throughput functional studies.

Key Results

  • Spatial Resolution: Achieved a physiological point spread function (PPSF) of 29 ± 0.8 µm laterally and 44 ± 1.6 µm axially in vitro, and 19 ± 3.7 µm laterally and 45 ± 6.1 µm axially in vivo.
  • Simultaneous Stimulation: Demonstrated simultaneous photostimulation of dozens of user-selected neurons in the intact mouse brain in vivo.
  • Volumetric Mapping: Successfully mapped synaptic connectivity in 3D volumes up to 800 × 800 × 200 µm³.
  • Speed: Achieved volumetric pattern refresh rates of 1.3 kHz, significantly faster than conventional CGH.
  • Cost Comparison: The 3D-MAP system costs approximately 23% of a high-performance two-photon system.
  • Minimal Cross-talk: Demonstrated minimal cross-talk between photostimulated ensembles in in vivo experiments.

Limitations and Future Work

  • Depth Penetration: Like other one-photon techniques, 3D-MAP's resolution is limited by light scattering in brain tissue, restricting its effective use to superficial layers (hundreds of microns).
  • Future Work: Suggests using red-shifted opsins and longer wavelengths to mitigate scattering. For deeper imaging, they mention surgical removal of overlying cortex or using miniaturized microscopes. They also propose pupil encoding, light field detection, or remote focusing to enable 3D detection over a larger axial range.
  • Sparse Labeling: The in vivo experiments relied on relatively sparse opsin expression for optimal specificity.
  • Future Work: For densely labeled populations, they suggest controlling stimulation pattern sparsity by multiplexing patterns to increase temporal sparsity.
  • Resolution Trade-offs: While 3D-MAP offers a good balance, it trades some spatial resolution compared to two-photon methods for lower power and cost.
  • Future Work: The authors suggest that for many experiments, the absolute spatial specificity of two-photon methods is not required, making 3D-MAP a preferable alternative.

Summary updated

Abstract

Optical control of neural ensemble activity is crucial for understanding brain function and disease, yet no technology can achieve optogenetic control of very large numbers of neurons at an extremely fast rate over a large volume. State-of-the-art multiphoton holographic optogenetics requires high-power illumination that only addresses relatively small populations of neurons in parallel. Conversely, one-photon holographic techniques can stimulate more neurons with two to three orders lower power, but with limited resolution or addressable volume. Perhaps most problematically, two-photon holographic optogenetic systems are extremely expensive and sophisticated which has precluded their broader adoption in the neuroscience community. To address this technical gap, we introduce a new one-photon light sculpting technique, three-dimensional multi-site random access photostimulation (3D-MAP), that overcomes these limitations by modulating light dynamically, both in the spatial and in the angular domain at multi-kHz rates. We use 3D-MAP to interrogate neural circuits in 3D and demonstrate simultaneous photostimulation and imaging of dozens of user-selected neurons in the intact mouse brain in vivo with high spatio-temporal resolution. 3D-MAP can be broadly adopted for high-throughput all-optical interrogation of brain circuits owing to its powerful combination of scale, speed, simplicity, and cost.
BibTeX
@inproceedings{xue2022three,
  author = {Yi Xue and Laura Waller and Hillel Adesnik and Nicolas C. Pégard},
  title = {{Three-dimensional multi-site random access photostimulation (3D-MAP)}},
  booktitle = {eLife},
  year = {2022},
  doi = {10.7554/elife.73266},
}

Loading related content...

Figure