A blog about neurophysiology

For quite some time, I was unsure about the reasons why images degrade when going to deeper layers with 2P point scanning. This also has remained largely unclear to me until the present point, after having done the estimates presented below, but at least now I’m getting a feeling for it.

The first difficulty lies in estimating whether the degradation comes from a degradation of the excitation PSF, or from low signal due to fluorescent photons that do not reach the detector due to scattering. The mean free path is around 200 um for the IR excitation light, and around 50 um for the to-be-detected visible light. Continue reading →

I have the impression that neurobiology is the single field in biology where many people with sometimes rather selective knowledge of the field have strong opinions about how it should work (i.e. how the brain should work). I don’t want to say that these opinions are worthless. In contrary, I enjoy them as an inspiring read that can question some of my beliefs that trickled down during the last years due to habituation. As a common sign, these theories present their new approach as a game-changer that offers a – until now – overlooked point of view on the brain. Nobody should be surprised by this, since, despite all efforts, an understanding of how specific information processing in the brain could look like is somewhat completely dark, and the most likely explanation for this lack of progress in understanding is a generally misguided paradigm in neurobiology, and the solution would be a paradigm change.

One example for a strong opinion I came across roughly two years ago, was the blog neuroelectrodynamics.blogspot.ch. (The author of the blog is apparently a professional neuroscientist.) On his blog, he highlights the importance of spike shape and propagation direction, and the possibility that these details of ion currents may lead to computations in neurons that are far more sophisticated than computations possible for simple firing rate models. It’s an interesting read, because it questions common places in neuroscience, although some of these opinions (firing rate models) are not necessarily part of the way most neuroscientists think of information processing anymore.

[Update 2018.] One example I came across recently is the blog mythsofvisionscience.wordpress.com, which questions the underlying assumptions that are driving the field of vision neuroscience (especially V1). It does not come up with a own crackpot theory of how the visual system should work, but instead points out many small and big weaknesses and flaws of current research in vision neuroscience. The blog is run by Lydia Maniatis.

Another (very different) example I found recently in the web is this homepage written in german: www.straktur.de It’s also available in english, but I have the feeling that it has been translated using Google Translate …
It stems from a mathematician who emphasizes the role of Glia cells; more precisely, he hypothesizes that Glia cells, responsible for nourishing neurons, evaluate the performance of neurons and support them accordingly. A dysfunctional neuron would therefore simply be discarded by cutting the support coming from Glia cells. By this, optimized information processing would arise naturally from this simple energy constraint (and this is where you can see the handwriting of a mathematician) – a basically interesting idea.
The author doesn’t really explain how the Glia cells might be able to evaluate the performance of surrounding neurons, but for simple circuits, one can imagine that this is possible.

At least for me, those kind of opinions are sometimes much more inspiring than a typical Nature/Neuron/Cell paper, because the presented opinion is strong, wants to convince he reader, questions the authorities in the field and often presents its theory as the holy grail, reminding you that there is something bigger out there (the holy grail) and that there is still a lack of understanding when it comes to specific information processing in neurons (or glia).

Genetically encoded calcium indicators (GECIs) are nowadays commonly used to report activity of many cells in transgenic animals; similarly, injected dyes like Rhod-2 can act as optical calcium reporters. The main shortcoming of this method is that it measures neuronal activity indirectly (via the calcium concentration), and, which is partially connected to the latter fact, at low temporal resolution (rise and decay times between 20 and 100 ms, or even more).

How nice would it be to have a genetically encoded fluorescent protein that changes its fluorescence not in response to calcium concentration, but to the electrophysiologically more relevant observable, the transmembrane voltage, better known as membrane potential. Indeed, these proteins exist. A number of recent papers made me aware of this fact; this week, I gave a short Journal Club about these indicators, and here I want to briefly summarize what is, according to my best knowledge, the state of the art.

Continue reading →

Granger causality has been named after the econometrician Clive Granger and has been adapted in the last 10-15 years as time-series analysis tool for neuroscience. The best account for this topic that I have found, is on scholarpedia again (link). The idea is quite simple: you have a timeseries (e.g. activity trace) X, and a timeseries Y. You want to know if the past of timeseries Y can, in addition to the past of timeseries X itself, help to predict the future of timeseries X. Prediction here is nothing but linear regression, somehow a mixture of auto- and cross-regression (copied from scholarpedia):

Continue reading →