17 November 2016

Why is neuroscience teaching software so bad?


Neuroscience is a discipline that is very well suited to using computer models. There are all sorts of elegant mathematical descriptions of how neurons generate action potentials, how signals propagate along the length of a neuron, how signals from neurons add up and contribute to firing action potentials or not, and more.

So why do so many pieces of software created to teach neuroscience suck so much?

Now let me say this: the teaching value of the software is often excellent. The problem is that the implementation is rough, twitchy, and out of date. So maybe my question is better phrased as, why does neuroscience teaching software suck in the context of using them today, in 2016?

Let me give a few examples. This year and last I’ve used Swimmy (Grisham et al. 2008). Students have to crack the neural circuit that makes a fish swim. It is an awesome exercise that challenges my students intellectually.

But when it starts...


You’re presented an MS-DOS command line. 1990s memory whiplash right there. The interface consists of lots and lots of windows you have to resize manually. And the Mac version is so out of date that it doesn’t run properly any more.

I tried some software at The Mind Project, including Virtual EEG (Miller et al. 2008). Virtual EEG has a cool and interesting premise. You can create different sets of pictures (say, photos of objects and photos of people), and the program shows averages of real EEG data that was generated by people viewing those pictures. But, again, the interface is kind of clunky and twitchy. It’s written in Java, and it still runs, but I ran into some refresh issues such that screens often didn’t refresh and display properly. It all worked, but was such a chore to get to the stuff I wanted.

Realizing that these efforts were done the better part of a decade ago, I went looking for mobile apps.

I only found one promising candidate, Neuronify (pictured above). This one, at first glance, seemed very promising. It runs on Android and iOS. The user interface is very clean. But it feels more like a neurophysiological sandbox for playing around than a teaching tool. You build stuff rather than being puzzles to solve. The commands are very limited. You can inject current into a cell, but you can’t specify by how much, for instance. I’m sure I could put it to good use, but I need to think about how to use it effectively.

Update, 3 April 2020: I have since learned that Neuronify has its own website, with desktop versions for Windows and Macs. The desktop versions also appear to have save options (important for being able to submit assignments in a teaching context), which I didn’t find in the app versions.

The contrast between teaching software and textbooks is profound. Textbook publishers have massive teams keeping the content and presentation up to date. There are new editions every few years. Say what you will about the cost, nobody would deny the typical university textbook is a professional looking, polished document.

Compared to the effort that goes into textbooks, most teaching software feels like the equivalent of a bunch of photocopied pages, printed off an old dot matrix printer, stapled together. It’s done by a small team, done once for some fairly specific teaching purpose, and nobody invests any effort in keeping up to date after it’s out and some small paper in an educational journal is published. So even those of us who decide to use the software have to pay the pixel tax.

I want students to struggle. But I want students to struggle with the inherent complexities of cellular neuroscience. Students don’t type in command lines to run Pokémon Go or Snapchat, and I don’t want them struggling with command lines in class. It’s the least important thing of all.

References

Grisham W, Schottler NA, Krasne FB. 2008. SWIMMY: Free software for teaching neurophysiology of neuronal circuits. Journal of Undergraduate Neuroscience Education 7(1): A1-A. https://mdcune.psych.ucla.edu/modules/swimmy/swimmy-extras/grisham-etal-junef2008.pdf

Miller BR, Troyer M, Busey T. 2008. Virtual EEG: A software-based electroencephalogram designed for undergraduate neuroscience-related courses. The Journal of Undergraduate Neuroscience
Education 7(1): A19-A25. http://june.funfaculty.org/index.php/june/article/viewFile/629/628

External links

Swimmy
The Mind Project (including Virtual EEG)
Neuronify
A pixel artist renounces pixel art

No comments: