Fundamentals Series: Deep-tissue optical recording and stimulation in behaving animals – An Introduction
In Vivo Multi-photon images, GCaMP6 in mouse hippocampus, using the MCI Pryer Endoscopic lens
Fundamentals Series: Deep-tissue optical recording and stimulation in behaving animals – An Introduction
Our next topic of the Fundamentals series will focus on a field with many exciting recent advances: deep-tissue optical recording and stimulation in behaving animals. We will discuss several popular methodologies for approaching this, as well as refer to some commercial solutions currently available.
This first chapter will serve as an introduction to the more method-focused chapters to follow. We also list the typical challenges in developing these technologies and in performing these types of experiments.
Recent progress in optical methods
Developments in genetics, synthetic biology and engineering technologies have fed off each other in recent years to make the goal of observing and optically controlling activity in the brain a reality.
Parallel advances in genetically-encoded reporters with high signal-to-noise ratios and optogenetics have been much publicized. These developments have been complemented by more affordable and sophisticated electronic architectures for other industrial or scientific applications (like CMOS and other detectors, fiber optics and other optical components).
In addition to this, some patent restrictions have been lifted on technologies such as multi-photon microscopy, and work-arounds discovered for others (e.g. the UCLA miniscopes project as an alternative to other commercial options). These have motivated more entrants attempting new designs.
This has had a further knock-on effect to entice part manufacturers to customise parts for biology applications, such as GRIN lens technology.
Why go deep, in-vivo?
The trend in research focus over the last decade has been from the previously more popular in vitro approach towards in vivo, where the intact network effects and relation to behaviour can be observed. This trend is demonstrated by publication keywords statistics (Google Scholar and PubMed), and the popularity of research papers on the topic.
This has been, to a large extent, helped along by the developments mentioned in the introduction. Electrical recordings and drug application in vivo have benefited from a similar development path to optical recordings, which acted as a further catalyst for in vivo optical technologies to be used more widely in neuroscience research.
Shifting towards optical, away from electric recordings
For many in the field, the experimental dream and the driving force is toward all-optical recordings in deep brain areas of a behaving animal in large 3-D volume. Add to this the ability to perform these recordings chronically and with high spatial and temporal resolution, ideally down to single dendritic scale, without the probe disrupting or interfering with behaviour. Before recent developments in the field moved our abilities closer to this ideal, using electrodes was the closest alternative to observing the brain in vivo.
The shift away from electrical control and recording is due to gene-targeting being better exploited by optical methods, both with recordings (genetically expressed calcium and voltage indicators) and control (optogenetic methods). It also offers the benefit of imaging less invasively in a larger volume, because of light’s ability to travel through tissue. Light can also be multiplexed in a way for the same probe to deliver and detect different colours, thereby custom-designing the experiment to address extra questions.
Of course, these are all potential benefits, and depending on the method employed for achieving the optical recordings. Besides recording technologies improving, co-expression of reporters and actuators is becoming more sophisticated. The focus is on reducing the excitation/activation bandwidth to prevent cross-talk and shifting these wavelengths towards longer ones for deeper penetration and less scattering through tissue. Currently, these opportunities are still also very prominent challenges with all-optical approaches.
Challenges with optical control and recording in vivo
Besides the issues of cross-talk and scattering with all-optical recordings, other important considerations centre around the following issues with in vivo applications, with most of these issues having some direct dependence on some others:
Always a major factor on many levels, concerns about invasiveness of the recording include tissue damage which may affect the optical data, but also damage which may impact wider brain function and normal behaviour. Much of the impact typically originates from the surgery. This may be the amount of tissue removed and how the tissue and the probe and/or animal is stabilised for recovery after surgery, if necessary.
1.1. How the probe is positioned – implant or not?
The type and position of the probe impacts the invasiveness of the method. For more superficial areas of the brain, data can be recorded without requiring an implant or removal of brain tissue. Current methods range from the less invasive thinning of the skull to the level where cortical emission can be detected through the thinned area, to a cranial window with removal of the bone, the dura and sometimes aspiration of some tissue. This choice is determined by both your target area as well as the strength of the optical signal and the spatial resolution required.
Implanted probes often have the benefit of allowing the animal to behave more freely, where recording optics attached to a microscope (multi-photon, typically) require a behavioural setup where the animal is restrained, like on a treadmill/virtual reality setup and its variations.
The non-implant solutions have the additional benefit of not being hampered by many other issues with implants, but they’re limited in terms of probing depth, with 2-photon solutions only allowing detection of cell bodies less than 1mm below the surface (ref).
Implants are more limiting in other ways, besides the obvious implications in terms of tissue damage. Focus can usually not be adjusted as easily, and in the case of some chronic recordings, the probe needs to be removed after the experiment and reinserted before the next. This exchange is restricted by the scope of an implanted guide cannula (ref). Achieving repeatability in targeting the exact same region between exchanges can be challenging.
1.2. Phototoxicity and photobleaching
More tissue between the probe light exit point and the target area requires an increase in light intensity. To achieve suitable signal, this may have to be increased to the level of affecting the physiology of the tissue.
Photobleaching of fluorophores can be particularly pronounced in vivo for exactly this reason, since exciting/activating the target area must be achieved in a volume of tissue.
3. Weight of the probe
With implants, the animal can typically carry around 10% of its weight on a head-mount, with adequate training (ref).
The way this weight is distributed can affect animal behaviour, for example a miniscope tower with high center of gravity). The location of the cranial window can also play a role, for instance having the mount balanced on top of the skull vs on the side).
4. Tethered systems
With freely moving animal experiments, several wireless systems have been commercialized in recent years. This will become a common feature with optical recording systems too, as is already evident with commercial systems for optical stimulation and recent publications evaluating a method for wireless photometry.
Commercially available systems for combined optical recording and stimulation using implants are still limited to optical or electric cables. The constraints of the tethers are usually reduced by incorporating rotary joints or commutators. This reduces cables winding up when animals move around.
With certain behavioural questions, especially those related to social interaction or in complex environments where there are many obstacles in the cage, wires and cables leading from the probe to the data acquisition box are particularly restrictive. With social interaction studies in mice, tethers are prone to be chewed, resulting in damage to the system and data losses.
5. Thermal Load
Circuitry and LEDs on the head and close to the probe often generate heat. This can be due to inefficient circuit designs, which will certainly improve with redesigns. It’s more difficult to improve on this front when there’s a requirement for dense processing electronics on the head, or high intensity LEDs, necessary when expression levels of the reporter/actuator are low.
Heat generation can be reduced by including some cooling infrastructure into the device. If not handled adequately, heating can alter physiological processes and behaviour of the animal. It may even damage tissue around the target area or just below the head-mounted electronics.
6. Data handling
Data handling is of critical importance, since more data has implications in terms of requiring thicker cables (both electric and optical cables). In the case of wireless systems, more data transfer typically translates into increased weight of the system. This is caused by the need for more battery power, more powerful on-board digitizers and more electronics in transmitter architecture. This factor can also have a knock-on increase on thermal load, since more processing may create more heat.
There are various strategies for achieving flexibility in data handling, which often include a compromise in the density of the recorded data, either with time, spatial resolution or reduction of field of view. Efficiency on this front is a core aspect of developing new systems (ref).
7. Location of the target area/cells
With implants, deeper structures mean more tissue damage. Damage is quadratically proportional to the diameter of the implant probe or the guide cannula (ref). It often requires removal of tissue before the implant can happen, with can alter network physiology.
8. Extra volume/depth
The fact that the experimenter is recording in extra volume, means that light scattering is an important consideration. Out-of-focus fluorescence can originate from the whole volume of tissue around the probe, which is a factor whenever the imaging method is not equipped with image sectioning components, like multi-photon, confocal or structured illumination.
There is no option for trans-illumination (at least not for animals rodent-sized and larger), which means contrast enhancement typically relies on epifluorescence and other methods.
9. Movement artefacts
In behaving animals, reducing movement artefacts is critical. Care must be taken with the design and implementation of the probe on the head, to offer as much stability as possible. Besides breathing and heartbeat, behaving animals impact the probe when socializing or interacting with obstacles in the home cage. Resulting shift of the probe’s focal plane and the difficulty in adjusting focus with implants can affect the recording dramatically.
A further method of handling movement artefacts is to express a reference/control reporter. This channel can then be detected as a deductible baseline for the effect of movement on the recording (ref). Movement correction algorithms (some examples on ImageJ/Fiji) may also be implemented during online or offline software analysis.
10. Bending stiffness
Closely tied to movement artefacts is the bending stiffness of implanted probes. The aim is to reduce the bending stiffness to the level where the target area can be reached accurately during implantation while damaging the tissue as little as possible when the probe moves during animal behaviour (ref). This variable is determined by the probe material, thickness of optical fibers and the substrate used to connect to the probe.
11. Working distance
Implanted lenses (GRIN lenses or implanted mini-lenses), which are currently the only option for reaching areas too deep for multi-photon setups, typically have a low working distance. To achieve focus without damaging the tissue therefore requires care and planning.
In the case of multi-photon recordings with restrained animals, the working distance of the lens determines how deep below the surface of the skull/cortex (depending on the surgery) the recording can be performed. The large optics manufacturers have a constant drive to increase this specification without compromising on other qualities.
12. Settling time for displaced tissue
For implanted probes and major tissue aspirations, a certain settling or recovery time is often required for the tissue to settle back to its supposed position around the implanted probe and return to full-functioning physiological activity. This can be challenging in certain experiments, where there may be uncertainty about whether the target area has been reached, only to find that this has not been the case after the settling time has expired.
In future chapters we’ll cover…
In this Fundamentals series, we’ll briefly address several technologies which are currently widely used, or which is being widely considered pending wider adoption by the community and positive evaluation by early adopters. These include:
- Fiber photometry
- Optical fiberscopes
- High-resolution Endoscopic lenses
These topics will be discussed in every subsequent chapter of the Fundamentals series, starting off here with fiber photometry.
Talk to our applications team about your in vivo optical recording and stimulation requirements – we have several interesting options available