In the class I'm teaching this semester on controversies in neuroscience, we had earlier talked about the conflict between Golgi and Ramón y Cajal regarding the nature of the neuron. Cajal argued that neurons are independent units while Golgi argued that they formed a continuous reticulum. One thing I wanted to emphasize to my students was that even though in hindsight we know that Cajal was mostly right and Golgi mostly wrong, based on the available technology at the time, both interpretations are highly possible since the data itself is very ambiguous. Also, much of how we interpret data is biased by what we know. So in Golgi's case, the reticular theory was what was accepted, and he was able to find what seemed like convincing evidence for it. As an exercise, I wanted to see if the students could replicate some of Cajal's findings. So I asked around my department for Golgi stained rat brain sections (which were hard to find, most people don't use this technique much anymore) and borrowed a microscope. I asked the students to look through the slides, find what they thought was unambiguously a single neuron, draw it to the best of their abilities and identify key features, such as dendrites, spines, soma, axon, etc. As you will see, some of the results are pretty interesting. Here's a picture of a Golgi stained portion of the cerebral cortex. This is not the section that was used, but it looked similar to this. Notice that while you can see a bunch of neurons, it is hard to tell whether they are separate or form a continuous web. One way to tease this apart is to show that at points where the processes cross, different ones are at different planes of focus:

Golgi stained section of the cerebral cortex of a rodent. Source: Neurodigitech
Now here's a drawing by Cajal of the Cortex. It is important to note that Cajal's drawings are not exact copies of his slides, but rather a synthesis of the various elements arranged in a way which would increase the clarity and logic of the way the data is represented. In fact, apparently many of his drawings were done by memory, after spending a day looking through the microscope, and done while having a drink in a local brothel. I know I have sometimes analyzed data at a coffee shop, but Cajal brings this to a new level.

Cajal's drawing of the cerebral cortex.
I think that one of the main errors committed by the students, was that if you look in many textbooks, neurons tend to look like this:

Diagram of a Neuron. Source: Wikipedia.
While this diagram helps to distinguish the different neuronal elements, the stumpy little dendrites, the gigantic cell body (soma) and the really thick axon somewhat mis-represent the actual proportions of a real neuron. For example, take a look at these two neurons drawn by students in my class (click on any of the drawings to enlarge):


Both students suggested that the longest, thickest process was the axon, while the skinny little ones were dendrites. In reality, what they probably thought was the axon was the apical dendrite, which is a very large primary dendrite which many cortical neurons have. In these cases the drawings were largely guided, I think, by a preconceived idea the students had of what a neuron should look like.
In this next drawing, the processes (axons and dendrites) are a bit more realistic, although soma seems a bit exaggerated. This student also assumes that the longest process is the axon, despite it having clear dendritic spines. Interestingly, the little smooth process near the bottom looks like the actual axon, or a stump of it, the rest was likely sliced off when the brain section was made:

This other student managed to guess that the little stump he drew near the bottom was the severed axon initial segment, althoug like the previous drawing the soma is exaggerated and there's a strange blob at the top of a dendrite, probably the out-of-focus soma of a nearby cell:

The best two I think are these next ones, they look more realistic, capturing the thickness of the dendrites relative to the soma, one correctly identified the stump of the axon, the other couldn't find an axon, but made no assumptions about it:


The truth is is that neurons are three dimensional, and often will have parts of their axons and dendrites cut off in a given brain section. That is why one needs to look at a lot of cells, and draw inferences from each one, finally putting everything together in a composite picture.
Overall, I think everyone enjoyed the exercise and it gave them a feel for what doing some of the experimental work was like. They also liked the little art school-style critique I did in class of their drawings. All of them did a great job and it was great fun. Next up: stimulus intensity coding in cockroach leg nerves...