Exploring the Concert of Neuronal Activities

November 02, 2010

How do real neural networks, composed of numerous different types of neurons, interconnected by complex arrangements of synapses, process information? Randy M. Bruno, Ph.D., Assistant Professor at the Department of Neuroscience, Columbia University, NY, USA is pursuing this question using the rodent whisker-barrel system. Here, anatomically and functionally distinct networks – barrels and barrel columns – are clearly identifiable, and the sensory transducers that provide input are directly controllable. With a variety of paired-recording techniques he investigates the mechanism for propagating information between thalamus and cortex, to study receptive field generation in excitatory and inhibitory neurons, and to demonstrate micro-organization of inputs to cortical columns. Using imaging techniques such as confocal and two-photon microscopy, Prof. Bruno visualizes neuronal dendritic arborization of neurons and their synaptic interconnections.

Fig. 1: Confocal Microscopy mosaic of a layer 4 in neuron filled in the barrel cortex of a living rat
Neuron at high resolution. Mosaic microscopy image recorded with Leica TCS SP5 confocal microscope. Filled neuron in the barrel cortex of a living rat.

What is the fundamental question in brain research?

We want to understand cortical circuitry – to know how this one circuit, iterated over the entire neocortex, solves tactical, visual, and cognitive problems. The outcome of many laboratories’ research is that you have the same cell types, arranged in the same laminar structures, and having the same general connectivity with each other and with other areas of the brain. It is as if nature reiterated this one circuit for many different tasks. Our goal is to reverse engineer that circuit.

Physiologists routinely record activity from individual neurons or groups of neurons to assess what the neuronal population is doing. But we all become anatomists in the process of doing this because we need to know how the neurons are connected, too.

We can use conventional tracers or newer methodologies like viral expression of fluorescence protein, label large groups of anatomical connections. And, in the course of doing the single cell recordings, we can label single axons. As we start to look at pairs of neurons, we’re trying to figure out the connectivity between individual cell types, or two particular cells, and get back to what the real circuit is.

Model organisms for investigation of different connectivities

Fig. 2: Several layer 4 cortical neurons (red) were filled in the barrel cortex of the living rat, and the barrels (green) subsequently stained by immunohisto-chemistry; scale bar: 300 µm
Two color image showing a cross section of the barrel cortex in a living rat, the barrels in green color and filled cortical neurons in red color. Images taken with Leica confocal TCS SP5 microscope.

To study the barrel cortex we work with rat and mouse. These two common laboratory species rely heavily on their senses of touch and smell because they are nocturnally active. Rats have this very stereotypical pattern of whiskers on their faces, which they use for tactile sensation the way humans use finger tips. They swipe their whiskers back and forth over objects and textures as they explore their environments, and they do it with the same frequency of palpation that humans use when we stroke our finger tips across something.

They have similar psycho-physical thresholds, so they can discriminate surfaces a little better than humans can, but they are basically very similar. The information from whiskers is processed by the barrel cortex. Barrels are very easily identified anatomical structures in the cortex: each barrel, a group of thousands of neurons, maps on to one whisker. So now we have a discrete sensory organ that we can control – a whisker – and an identifiable network that is listening to it. We can control the input and take apart the network.

We use electrophysiology approaches on anesthetized, sedated, and conscious head-fixed animals. We are now getting into behavioral studies because ultimately sensation is an active process.

Fig. 3: High magnification view of a short segment of dendrite shows numerous spine heads, the sites of synaptic connections; scale bars: 3 µm
Maximum Intensity projection of a sequence of sections through a short segment of dentdrite showing numerous spine heads. High resolution dendrite imaging with Leica TCS SP5 confocal microscope.

Technical approaches

Everything we do is in vivo, although we are now starting to work in slices. We heavily rely on whole cell recording in vivo to actually patch into neurons and record intracellular membrane potential as well as action potentials. This approach is wonderful for looking at synaptic inputs and is key to the research.

We also use a lot of conventional physiology recording techniques like extracellular recording of single units and local field potentials. We do two-photon imaging of both voltage and calcium sensors. We also do a lot of anatomy, looking at the dendrites and axons of single cells we’ve recorded from. For examining large axonal tracks, we employ conventional tracers and viral mediated expression of GFP and many fluorophores.

For anatomical purposes, we use broadband confocal. On fixed tissues we image either GFP or dyes like Alexa. A custom two-photon microscope is for anesthetized and conscious in vivo experiments where we use a variety of synthetic dyes to measure voltage or calcium. We are also experimenting with different viruses in the lab for expressing different genetically encoded indicators.

Fig. 4: Examples of putative synaptic contacts
Putative synaptic contacts discovered by filling two cortical neurons with two different colors and imaging with two-photon excitation deep inside the living brain.

Technical limits

I have never met a scientist who is completely happy with the technologies that are available. So, yes there are limitations. What we can do today is fabulous, but I think we are almost insatiable when it comes to technology. So when we use confocal imaging of fixed tissue to map out structures and detailed morphology, we don’t image for the purposes of getting nice pictures.

We’re usually trying to obtain something we can quantify, and that often means that we scan large structures (hundreds of microns) in 3D, but we have limited resolution due to diffraction limits. Regarding the diffraction limit and the depth of recording, new technologies such as STED and the OPO laser can contribute toward solving these problems. But they probably cannot overcome all limitations. One field for new development is better dyes, especially with regard to sensitivity. These are problems that we desperately need molecular biologists and organic chemists to overcome.

To do good neuroscience these days, you have to frequently combine incredibly different skill sets. You need computer programming, molecular biology, physiology, anatomy, physics, and chemistry, just to mention the most important ones. That being said, this is really fun and exciting because there are so many distinct skill sets involved and so many people to collaborate with.

Comments