I somehow missed that Josh Vogelstein’s method on action potential detection was published last summer. In Spike Inference from Calcium Imaging Using Sequential Monte Carlo Methods, the authors use a Monte Carlo approach to determine spike times from calcium imaging with superior performance to other deconvolution methods. It does a great job on simulated and in vitro data, I’d love to see performance on real in vivo recordings. If you are serious about calcium imaging, you should definitely get in touch with Josh and see what magic he can do with all that math. You should also ask him about the benefits of linen pants vs. denim, he’s got strong opinions on that subject as well…
Using only strongly saturating and very noisy in vitro fluorescence measurements to infer precise spike times in a ‘‘naturalistic’’ spike train recorded in vitro
The Deisseroth and Hegemann groups have just published a newly engineered channelrhodopsin, ChETA, in the Nature Neuroscience paper, Ultrafast optogenetic control. Gunaydin et al rationally targeted mutations to the opsin pocket of channelrhodopsin-2 to increase the speed of channel deactivation/closing. ChETA provides higher fidelity optical control of spiking at high expression levels or firing frequencies (up to 200Hz!) and eliminates plateau potentials during sustained spike trains.
ChETA improves spike train fidelity
ChETA clearly provides higher precision in optical control of spiking, particularly at high spike rates. However, a big problem limiting some in vivo channelrhodopsin use has been insufficient conductance. Some groups have sought to increase single channel conductance, but this approach can lead to increased ChR toxicity and/or spurious spikes. At first glance, increasing deactivation rates, and thus decreasing single channel current from a brief light pulse, seems to make life MORE difficult for situations where light and conductance levels are limiting. ChETA produces a very low number of successful spikes at 1ms illumination (Fig 3f), as compared to ChR2. ChETA response peaks within 1ms but requires 2ms illumination and >10Hz trains to induces spikes more reliably than ChR2. Is this due to a decreased peak single channel conductance in ChETA or just the activation/deactivation rate differences? I couldn’t find a direct single-channel conductance comparison in the paper.
ChETA requires a longer light pulse than ChR2 to generate spikes.
A reduced conductance might not be a bad thing though. ChETA’s increased deactivation rate might make less toxic to cells, allowing a higher expression level, which would compensate for a reduced single channel current flow. It all depends on what causes ChR2 toxicity. Is toxicity caused by a non-illuminated leak current or something else? Is the deactivation rate correlated with a leak current and/or toxicity? I would love to see a quantitative comparison of expression level and toxicity between wt ChR2 and ChETA. Maybe our readers can post their experiences with it in the coming weeks.
Imaging fluorescence in an intact, living brain is difficult due to absorption and scattering of excitation and emission light. Two photon microscopy uses excitation light in the narrow optical window (700-950nm) where water and hemoglobin do not significantly absorb, which allows structure determination and functional imaging down to depths of ~600nm from the surface of the brain. However, scattering of the excitation light still occurs at these wavelengths, which distorts the excitation volume and causes a rapidly increasing fluorescent background at greater depths.
The vasculature was labeled by injecting flourescein dextran into the circulatory stream. The light source was a regenerative amplifier. ‘‘0 mm’’ corresponds to the top of the brain. Left, XZ projection. Right, examples of XY projections. Note the increase in background fluo- rescence deeper than 600 mm in the brain due to out-of-focus 2PE. (Theer et al., 2003)
In order to reduce this background and sharpen the two-photon excitation volume, Ji et al in Adaptive optics via pupil segmentation for high-resolution imaging in biological tissues adapt tricks from astronomy (see also Rueckel et al 2006). Depending on where each ray of excitation light enters the brain, the angles of scattering are different. By sequentially illuminating with spatially restricted subsections of full illumination beam, they see how the brain warps the light from each part of the beam. Then they use a flexible mirror and to change the angle and phase of the excitation beam so that each part of the beam comes together in the same place and phase. Since two-photon fluorescence scales as the square of the intensity of the excitation light, this dramatically improves both the resolution and the signal to noise of the fluorescence image.
Scattering in the brain warps two-photon excitation light, but adaptive optics can correct this.
Pre-warping the excitation light to cancel the scattering improves fluorescence localization and intensity.
How well will this work in the brain of a living animal? It’s not clear. Large differences in scattering paths across a field of view may dramatically slow the determination of optimal excitation beam warping and make it difficult to scan across a field of view quickly. Motion of the brain may change the pattern of scattering faster than the system can adapt. Still, the use of adaptive optics is one of the few promising techniques for increasing penetration depth and signal in optical brain imaging (not counting sucking off the bits of the brain that are “in the way” of the excitation beam.) There are other optical tricks in astronomy that have not yet been mined for neuroscience applications. Hopefully these will one day allow the functional imaging of all layers of the mammalian cortex, not just layer 2/3.
Ed Boyden’s group has published High-performance genetically targetable optical neural silencing by light-driven proton pumps, detailing a set of new optical neuronal off-switches borrowed from various species that appear to be much better than Halorhodopsin for silencing neurons. Halo works well for preventing action potentials when the nucleus is illuminated, but has a harder time blocking transmission of action potentials down an axon after it has been initiated. Also, previously engineered Halo variants, including eNpHR, suffer from light-dependent inactivation and have an expression sweet spot that could use broadening. I’m looking forward to finding out if Arch, and the other new switches from Boyden’s group allow more powerful experiments in our hands.
In two examples of imaging from a mouse ear, (above) shows the distribution of TBO, a photodynamic therapy drug, following drug administration, (below, red) shows the distribution of hemoglobin in blood vessles.
How specific is the detection of endogenous chromophores? They report that 60nm is the absolute detection limit, but this is for a pure chromophore in water. In real cells there will be many other endogenous chromophores at various concentrations. For example, endogenous background fluorescence of flavins is often easily seen when imaging at CFP wavelenghts. Watt Webb has been imaging those types of chromophores for years. The intersection of both a preferred excitation wavelengh and a preferred stimulated emission wavelength will provide some selectivity, but I suspect this will be most useful for imaging the distribution of fairly highly expressed chromophores in vivo. Distinguishing chromophores with highly overlapping spectra may not be possible. Of course, many, many proteins don’t have distinctive chromophores (tyrosine does not count!) built in to them, so GFP won’t be out of work any time soon. However, this stimulated emission imaging doesn’t require transgenic or small molecule labeling, so it could potentially allow imaging in humans.
Do the absorbance and emission spectra and the excited state lifetime provide sufficient selectivity to detect low concentrations of chromophores in vivo?
Three papers are out online in Nature Methods that show big improvements in calcium imaging with genetically encoded sensors. They are are based on the fluorescence intensity indicator, GCaMP. GCaMP, first developed by Junichi Nakai, consists of a GFP that has been circularly permuted so that the N and C termini are fused and new termini are made in the middle of the protein. Fused to one terminus is calmodulin and the other is a peptide, M13, that calmodulin (CaM) binds to in the presence of calcium. The name is supposed to look like GFP with a CaM inserted into it, G-CaM-P. Normally the GFP is dim, as there is a hole from the outside of its barrel into the chromophore. Upon binding calcium, this hole is plugged and fluorescence increases.
The first paper, A genetically encoded reporter of synaptic activity in vivo, from Leon Lagnado’s group, targets GCaMP2 to the outer surface of synaptic vesicles. This localization allows the fluorescence signal to be confined to the presynaptic terminal, where calcium fluxes in response to action potentials are high. This targeting improves the response magnitude of GCaMP2 and permits the optical recording of synaptic inputs into whatever region of the brain one looks at. They demonstrate the technique in live zebrafish.
In the second paper, Optical interrogation of neural circuits in Caenorhabditis elegans, from Sharad Ramanathan’s group, GCaMP2 has been combined with Channelrhodopsin-2 to perform functional circuit mapping in the worm. Since the worm’s structural wiring diagram has been essentially solved, functional data could say much about how “thick” the wires between each cell are. Unfortunately, with GCaMP2, the responses are too slow and weak to distinguish direct from indirect connections.
Finally, we have published a paper, Imaging neural activity in worms, flies and mice with improved GCaMP calcium indicators, describing the improved GCaMP3. This indicator has between 2-10x better signal to noise than GCaMP2, D3cpv and TN-XXL, depending on the system you are using. It’s kinetics are faster and it is more photostable than FRET indicators, and the responses are huge. When expressed in motor cortex of the mouse, neuronal activity is easily seen directly in the raw data. Furthermore, the sensor can be expressed stably for months, making it a potential tool for observing how learning reshapes the patterns of activity in the cortex.
Imaging of mouse motor cortex (M1) expressing the genetically-encoded calcium indicator GCaMP3 through a cortical window. After 72 days of GCaMP3 expression, large fluorescence transients can be seen in many neurons that are highly correlated with mouse running.
GCaMP3 is not perfect. It cannot reliably detect single action potential in vivo in mammals, though I doubt that any existing GECI can. Work continues on future generations of GCaMP that may achieve 100% fidelity in optical reading of the bits in the brain. However, there is considerable evidence from a number of groups that have been beta-testing the sensor, including the Tank lab of “quake mouse” fame, that it is a significant leap forward and unlocks much of the fantastic and fantasized potential of genetically-encoded calcium indicators.
Comparison of fluorescence changes in response to trains of action potentials in acute cortical slices.
I will try to post a more complete writeup of GCaMP3 for Brain Windows soon, with an unbiased eye to its strengths and weaknesses. We worked very hard to carefully characterize this sensor’s effects on cellular and circuit properties. If you have any questions about GCaMP3, please post them to the comments.
Most people play Quake with a computer mouse, but researchers in David Tank’s lab at Princeton have done it with a living mouse, AND they are recording the intracellular activity of individual neurons of the mouse during the gaming session. As reported in Intracellular dynamics of hippocampal place cells during virtual navigation, the virtual reality environment of the video game was sufficiently realistic to generate place cell activity in the mouse’s hippocampus.
Now where did I see that cheese power-up?
Place cells modulate their activity dependent on the location the mouse is at. They have mostly been identified with extracellular recordings in freely moving mice. Extracellular recording only permits the detection of the rates of action potential firing, rather then the subtle intracellular voltage changes that could help explain the mechanism of place cell activity generation. A few pioneers, such as Albert “my greatest strength is a tremendous capacity for boredom” Lee, have recorded intracellularly in freely moving animals, but these experiments are fiendishly difficult, as the motion of the animal’s head tends to break the seal on the recorded neuron. Only a few cells have been recorded in that manner for more than a few minutes, though the success rate has been improving recently.
Experimental setup
In Chris Harvey’s technique, they fix the head of the mouse to a bar and let the mouse walk on a floating ball, while a virtual reality screen is projected in the mouse’s field of view. The motion of the ball controls the motion on the screen. The head never moves, so intracellular recordings can be made relatively easily and held for long periods of time.
The authors find three characteristics of place cell activity that could explain their generation and function.
“An asymmetric ramp-like depolarization of the baseline membrane potential, an increase in the amplitude of intracellular theta oscillations, and a phase precession of the intracellular theta oscillation relative to the extracellularly recorded theta rhythm.”
Intracellular voltage dynamics in place cells
These could be used to explain how place cells remap their selectivity when a mouse (or a human) moves into a new environment. This also could be used to do more in depth studies of the mental replay of place locations that has been previously recorded in the activity patterns of the hippocampus. The technique itself is about as sexy as neuroscience gets. Unfortunately, this paper also provides an additional piece of evidence for Karel to use in motivating lab post-docs, “Look at Chris, he left the lab after you got here and already has a Nature article…”‘
One of the most time consuming and frustrating tasks associated with fluorescence imaging in the brain is picking out your regions of interest. Which pixels do you include in as part of the cell and which are part of the surrounding neuropil? Often, the answer is not obvious, and even with painstaking selections you can make errors. Eran Mukamel et. al, from Mark Schnitzer’s lab just published this Neurotechnique Automated Analysis of Cellular Signals from Large-Scale Calcium Imaging Data that aims to simplify and improve the results of ROI selection.
The authors used a multistage approach to identify and quantify the calcium-dependent fluorescence changes of imaged neurons. First, they used principal component analysis to identify the components of the image that were likely calcium signal related and which were noise. The sparse nature of the calcium response (calcium transients are brief and spatially confined) helped the separation from the noise. They threw the noise away. Then they used independent component analysis to pick out which components of the calcium signal changed in a manner independent from other pieces of the signal. These likely represent individual cells. Using this output, they performed auto-segmentation of the image into numerous individual neurons or processes and measured the fluorescence change in those regions. In simulations of data, it resulted in superior data fidelity over hand drawing ROIs. They also validated it with real in vivo calcium imaging.
Automated Cell Sorting Identifies Neuronal and Glial Ca2+ Dynamics from Large-Scale Two-Photon Imaging Data
Whether its neuronal imaging, high-speed motion tracking or multielectrode recordings, tremendously large data sets are currently being generated in systems neuroscience. It is simply impossible for a single post-doc to crunch all of her data without major automated computational techniques. In calcium imaging, the resources that have been poured into the development and release of powerful new tools requires an equal effort on the data analysis end to maximize the value of this technique. The automated algorithms presented in this paper look very promising and we will definitely be checking them out in the near future.
Speaking of reactivating specific memories, at the 2009 Society for Neuroscience meeting, Matteo Rizzi of Michael Häusser’s lab is presenting the realization of an idea that has been floating around in some research proposals I’ve read over the last year. Express channelrhodopsin-2 under control of the immediate early gene c-fos, induce a strong memory formation via fear conditioning, and then drive the recall of that memory by stimulating the neurons that are expressing ChR2. Immediate early genes are activated shorty after high levels activity in neurons, though the precise patterns are different depending on which promoter (c-fos, Zif268, etc) you use, making precisely HOW they reflect recent neuronal activity patterns unclear. Nevertheless, the activation of the c-fos based pattern seems close enough to trigger an identical behavioral response as the conditioned stimulus.
Not yet, but getting closer...
Electrically-induced fear conditioning is probably the most blunt instrument possible, encoding a very powerful, general ‘fear’ memory, and many things can make a mouse freeze. Thus, this is definitely the low-hanging fruit on the ‘reverse-engineering’ memories tree. Understanding how the information in a memory is distributed across participating neurons is going to take a more sophisticated approach and a lot more work. This result is still incredibly cool, and I’m somewhat surprised it worked by driving ChR2 with c-fos in a hundred cells in the dentate gyrus. That has pretty powerful implications for avenues by which memories can be recalled. Surely the entire memory is not encoded by only the 100 neurons that were activated! How many other neurons participate, and how does the optical stimulation activate the entire ensemble? Is it even necessary to activate the entire ensemble to drive behavior? The poster will be MOBBED. I look forward to reading the details.
Program#/Poster#:
388.8/GG103
Title:
Memory recall driven by optical stimulation of functionally identified sub-populations of neurons
Location:
South Hall A
Presentation Time:
Monday, Oct 19, 2009, 10:00 AM -11:00 AM
Authors:
*M. RIZZI, K. POWELL, J. HEFENDEHL, A. FERNANDES, M. HAUSSER;
Wolfson Inst. for Biomed. Res., UCL, London, United Kingdom
Abstract:
The mammalian brain is capable of storing information in sparse populations of neurons encompassing several brain areas. Immediate recall of this information is possible upon presentation of a cue or context. Most aspects of this process remain unresolved: are the cells involved in information storage also responsible for its recall? What portion of this distributed circuit needs to be reactivated, in order to achieve successful recall? To answer these questions we selectively expressed a genetically encoded optogenetic probe (Boyden et al., 2005) in neurons engaged during the learning of a specific association. A plasmid encoding channelrhodopsin-2 and EGFP under an immediate early gene promoter (c-fos-ChR2-IRES-EGFP) was electroporated in vivo into granule cells (GCs) of the dorsal dentate gyrus of anaesthetized C57BL/6 mice. Mice were allowed to recover, and then underwent classical delay fear conditioning (consisting of 10-20 pairings of a 5 second auditory tone and a 2 second footshock). An optic fiber was implanted intra-cranially to allow optical stimulation of transfected neurons. Light stimulation (λ = 530 nm; 5 Hz) successfully induced recall of the fear memory, measured as freezing behaviour (n = 27 animals). Post-hoc analysis of the transfected tissue revealed that a remarkably small subpopulation of GCs (<~100 cells) was sufficient to cause this effect. We then tested whether any, comparatively sized, subset of GCs could be equally effective. We transfected neurons with a plasmid encoding ChR2 expression under a general promoter (pCAG-ChR2) to obtain ChR2 expression in a random population of cells. Interestingly, optical stimulation of this population was insufficient to induce memory recall (population data: n=30). Our results therefore suggest that recall of a learned association, sparsely stored in neuronal circuits distributed over several brain areas, can be achieved by the simple reactivation of a very small subset of neurons involved in learning this association. Furthermore, our strategy may also be useful for dissecting the complexities associated with memory storage and recall.
Support: Gatsby Charitable Foundation; Wellcome Trust
apo-PhyB covalently binds to the chromophore phycocyanobilin (PCB) to form a light-sensitive holoprotein. PhyB undergoes conformational changes between the Pr and Pfr states catalysed by red and infrared light, reversibly associating with the PIF domain only in the Pfr state. This heterodimerization interaction can be used to translocate a YFP-tagged PIF domain to PhyB tagged by mCherry and localized to the plasma membrane by the C-terminal CAAX motif of Kras.
When asked about the possibility that this could be used in-vivo, Levskaya said
The only real caveat for in-vivo work is delivery of the non-native PCB tetrapyrrole. From the literature and my experience with cell culture I suspect it shouldn’t be hard to just administer it directly to animals to get saturating levels for holoprotein formation. It might even be possible just to feed animals Spirulina (where it comes from). There’s nutrition literature that suggests their livers are capable of freeing PCB and getting it into the blood stream.
Observing light-induced Cdc42 activation with a TIRF recruitment biosensor
Expression of genetic tools that control neural activity (Channelrhodopsins, Halorhodopsins, DREADDs) in functionally defined populations, such as neurons that are active during a particular task or thought, is the next big leap that needs to be made in systems neuroscience. This may be achieved by combining an imaging technique to identify active neurons, such as G-CaMP3, with photo-switchable transcription. The technique presented in the above paper is one promising avenue which may lead to cell-specific photo-switchable transcription. Once robust versions of these tools are in place, scientists will begin to work out the complex and thrilling processes of reverse-engineering and manipulation of specific thoughts and memories, at least in mice and rats.
Recent Comments