Optical monitoring of neuronal activity at high frame rate with a digital random-access multiphoton (RAMP) microscope

Abstract

Two-photon microscopy offers the promise of monitoring brain activity at multiple locations within intact tissue. However, serial sampling of voxels has been difficult to reconcile with millisecond timescales characteristic of neuronal activity. This is due to the conflicting constraints of scanning speed and signal amplitude. The recent use of acousto-optic deflector scanning to implement random-access multiphoton microscopy (RAMP) potentially allows to preserve long illumination dwell times while sampling multiple points-of-interest at high rates. However, the real-life abilities of RAMP microscopy regarding sensitivity and phototoxicity issues, which have so far impeded prolonged optical recordings at high frame rates, have not been assessed. Here, we describe the design, implementation and characterisation of an optimised RAMP microscope. We demonstrate the application of the microscope by monitoring calcium transients in Purkinje cells and cortical pyramidal cell dendrites and spines. We quantify the illumination constraints imposed by phototoxicity and show that stable continuous high-rate recordings can be obtained. During these recordings the fluorescence signal is large enough to detect spikes with a temporal resolution limited only by the calcium dye dynamics, improving upon previous techniques by at least an order of magnitude.

Introduction

Monitoring cell type specific activity in space and time is central to our understanding of brain microcircuits function. Imaging techniques are well suited to yield large-scale non-invasive recordings of local microcircuit activity with cellular precision. In particular, specific cell types can be selectively targeted on the basis of their morphology or according to the expression of genetically encoded fluorescent markers and probes (Diez-Garcia et al., 2005, Hasan et al., 2004, Heim et al., 2007, Yaksi and Friedrich, 2006). Following the recent development of bulk loading protocols (Nagayama et al., 2007, Nevian and Helmchen, 2007, Nimmerjahn et al., 2004, Peterlin et al., 2000, Stosiek et al., 2003), optical recordings of neuronal activity have brought new insights into the functional organization of microcircuits in vitro (Brecht et al., 2004, Cossart et al., 2003, Ikegaya et al., 2004) and in vivo (Kerr et al., 2005, Nagayama et al., 2007, Ohki et al., 2005, Sullivan et al., 2005, Yaksi and Friedrich, 2006).

Monitoring single spikes at high temporal resolution in a population of neurons can be achieved with photodiode arrays or fast CCD cameras (Mao et al., 2001, Smetters et al., 1999), but these field illumination techniques suffer from serious drawbacks. Their spatial resolution, penetration capability in diffusive brain tissue and signal-to-noise ratio are poor. Moreover, high photon flux induces photodamage and photobleaching, limiting the duration of the recordings (Mao et al., 2001). Laser-scanning techniques, which have optical slicing capabilities and can resolve neurites in situ, could overcome these limitations. Non-linear excitation microscopy (Denk et al., 1990), in particular, has greatly contributed to improve the depth of penetration and the sensitivity of in situ imaging (Denk et al., 1994, Helmchen and Denk, 2005) at reduced photodamage levels. However, laser-scanning microscopes are serial-acquisition devices and their temporal resolution is limited both by the inertia of mechanical scanners and by the vanishingly small signal collected from each voxel at high scanning speed. Despite efforts to implement parallel scanning with multiple beams (Hell and Andresen, 2001, Straub et al., 2000) the use of laser-scanning imaging has been so far confined to the study of slow population activity (Garaschuk et al., 2006) like calcium waves (Crepel et al., 2007, Stosiek et al., 2003), cortical up-states (Kerr et al., 2005, Yaksi and Friedrich, 2006) or responses to prolonged stimuli (Ohki et al., 2005, Yaksi and Friedrich, 2006).

In laser-scanning microscopy, images constitute a bidimensional sampling of tridimensional biological objects. A simple strategy to improve the temporal resolution is to further reduce the dimensionality of the scanned area. Line scanning has been extensively used to follow biological signals with kHz resolution (Yasuda et al., 2004). The spatial sampling capacity of this technique has been recently adapted to the study of neuronal populations by implementing curved trajectories in 2D and in 3D (Gobel et al., 2007). However, one-dimensional scanning is not optimal to sample the sparse structure of neurons and networks. Ideally, points of interest should be placed on each singular object. This ultimate reduction in the dimension of the scanned area, aimed at optimizing useful dwell times, was named random-access microscopy (Bullen et al., 1997). Such discontinuous scanning cannot be implemented with galvanometric mirrors but is the essence of AODs, which are digital pointing devices with microsecond switch times (Bullen et al., 1997). Adaptation of AODs to two-photon microscopy has raised important technical difficulties that have been solved in the past years (Iyer et al., 2006, Iyer et al., 2003, Kremer et al., 2008, Lechleiter et al., 2002, Reddy and Saggau, 2005, Salomé et al., 2006, Zeng et al., 2006, Zeng et al., 2007) leading to the design of functional random-access multiphoton (RAMP) microscopes (Iyer et al., 2006, Salomé et al., 2006, Vucinic and Sejnowski, 2007).

RAMP microscopy has been employed in episodic acquisition protocols to monitor calcium influx at high speed for short-duration trials (Diana et al., 2007, Iyer et al., 2006). However, the compatibility of RAMP microscopy with continuous network imaging remains to be established, as repetitive two-photon excitation at high rates is known to cause photodamage (Hopt and Neher, 2001, Koester et al., 1999, Yasuda et al., 2004). In the present work we aimed at optimizing the sensitivity and reducing the phototoxicity of RAMP microscopy. We describe the design choices made to build a fully interfaced RAMP microscope in which the illumination and collection of light have been optimised for high-speed scanning. As preliminary to real-life experiments, we establish the laser power threshold for the phototoxicity of continuous high frame rate RAMP imaging. The sensitivity of our instrument allows us to obtain stable high-speed recordings without phototoxicity. We then demonstrate the microscope's application in two biological preparations: in cerebellar Purkinje cells, we detect complex spike calcium influx at single spines with millisecond precision; in neocortex, we show that the calcium transient evoked by simple spikes in apical dendrites can be detected. The temporal precision of single-event detection (3 ms) is limited only by the dynamics of the fluorescent dyes used, and is significantly improved compared to previous methods.