As the study of physiology moves to the cellular and molecular level, researchers need tools that support measurements with high spatial and fast temporal resolutions. The ability to measure cellular processes that change quickly, such as neuronal or cardiac action potentials in living tissue, is crucial to understanding cellular and molecular functions. We would like to be able to make measurements at many sites on a specimen simultaneously to study how single cells function or how networks of cells interact. Many current instruments, however, require tradeoffs between spatial and temporal resolution. Physical probes—such as microelectrodes—allow direct sensing of electrical signals in cells with high-speed sampling but limit the number of sites on a specimen that can be concurrently studied. Alternatively, indirect measurements can be made at many sites with optical imaging systems—by labeling cells with indicators that change their fluorescence intensity in response to parameters, such as membrane potential or intracellular ion concentrations—but most existing systems are relatively slow.1
Confocal microscopy allows imaging within light-scattering biological samples by combining point illumination with point detection to accomplish optical sectioning.2 Unfortunately, most available confocal microscopes must build up an image on a pixel-by-pixel basis and are limited to frame rates of about 1Hz. This is much too slow to record fast signals such as action potentials, which occur on a timescale of milliseconds. To increase the effective frame rate to about 500Hz, spatial information can be sacrificed by restricting the area studied to a single line instead of a full frame.3 This scheme, however, does not allow for measurements at user-selected sites distributed throughout the specimen: often, only two or three sites yield usable data with a single scan line. To overcome this limitation, we take advantage of the fact that most cellular specimens are inherently sparse. By focusing all of the recording time on sites of interest with an addressable confocal microscope, we maximize the spatio-temporal resolution.
To achieve high-speed addressable point illumination, we use acousto-optic deflectors to steer a focused laser beam to sites of interest. For high-speed addressable point detection, we use a digital micromirror device to create a virtual pinhole that tracks the fluorescence from the excited sites (see Figure 1).4 These devices can be updated to a new location on the specimen every 20μs. By allotting another 1–5μs for sampling, the system can achieve a pixel rate of about 45kHz. Depending on the temporal characteristics of the parameter under investigation, we can change the pixel rate to observe either many sites that produce slow signals or fewer sites that produce fast signals. Even with 10 sites of interest, our system can achieve an effective frame rate an order of magnitude faster than many line-scanning systems. Furthermore, we can dramatically improve the signal-to-noise ratio by oversampling at each site.
Figure 1. Acousto-optic deflectors (AODs) direct laser light to user-selected sites. The light excites fluorescence in the specimen. An addressable confocal pinhole implemented by a digital micromirror device (DMD) tracks the fluorescence and directs in-focus signals towards a photomultiplier tube (PMT).
To demonstrate the ability of the system to obtain images from light-scattering specimens, we collected structural images of green-fluorescent-protein-labeled neurons inside 350μm-thick mouse brain slices (see Figure 2). This slice thickness is sufficient to preserve the normal in-vivo cellular architecture of the neurons. When we reconstruct the data in 3D, we can clearly see fine details of the neuronal structure. The functional imaging properties of the system are demonstrated by recording calcium transients from a cultured neuron filled with Oregon Green BAPTA-1, a molecular indicator (see Figure 3). The transient signals are elicited by delivering electrical stimulation through a single probe at the cell body. From this data, we can see that the neuron's response is not uniform throughout, and differential signaling can be visualized at different sites of interest.
Figure 2. A 3D visualization shows the fine structures in fluorescently-labeled neurons in a 350μm-thick brain slice.
Figure 3. Action potentials (APs) were elicited with current injections at site ‘P’ in a neuron and measured electrically. Simultaneous intracellular calcium transients were optically monitored from several user-selected sites on the neuron (1–6).
We developed an addressable confocal microscope, using acousto-optic deflectors and a digital micromirror device, that provides the spatial and time resolution needed to study processes at the cellular and molecular level. This system will allow researchers from a wide range of disciplines to study novel aspects of cellular signaling that were previously undetectable due to the limitations of currently-available tools.