| Neutral | Another industry standard. Very difficult to use. Contains non-portable workflows. Forces you to change units every time you change scale (which may be a good thing for beginners, but it's very annoying when you need to get things done). Uses highly non-standard connection maps and generally is idiosyncratic and difficult to interface with other systems. Other than Annie which focuses on astrocytes, there are no known tools to generate Brian2 code. |
| Neutral | This isn't a fully functional simulator, rather it's an interesting take on the user interface. The user interface is reminiscent of others like Nengo, it's a nice way to create network topology interactively. |
| Thumbs Down | GENESIS used to be an industry standard but it isn't being maintained lately and it's been mostly superseded by NEURON in the field. It still works, but it gets a thumbs down on the basis of lack of maintenance and lack of currency. And like most other traditional simulators, it doesn't handle field potentials correctly. This probably goes into the "don't bother" category, if you're starting something new don't use this tool, it's too old and it won't do what you need. Unfortunately it wasn't built for interoperability in the first place, it has a lot of legacy ascii file formats that no one else can read. If you have existing simulations in this tool, my best advice is to get them out as soon as possible. One way is to export them as NeuroML, and then import them somewhere else. On the off chance that you have something you want to get "in" to GENESIS, this tool might help you, although it too is outdated and no longer being maintained. |
| Thumbs Up! | GeNN is a GPU enhanced neural network simulation environment. It works great with older GPU's like the nVidia GTX-1080. It has a Python wrapper called PyGeNN.It has its own network definition language and handles some oddball architectures that others have difficulty with. |
| Neutral | Old stochastic simulator for cell biology. There is a Blender 5.1 version, so it looks like it's being maintained. Has a version for Windows 11 (only, not 10), and requires Blender. |
| Neutral | A pretty nice Windows application but it hasn't been maintained in a few years and on that basis gets a neutral. It's an interesting program in that it generates C code, which is good for a lot of things. |
| Thumbs Up! | Installs perfectly on Windows, works right out of the box. Very nice simulator tool, the user interface works in a browser and you can build a network with one or two lines of Python code. Nengo does some unusual things with the i-F curves that are well worth looking into, the updating method can be applied to other parameters to achieve a full spectrum of plasticity. |
| Thumbs Up! | BlueBrain work product. Works great in Docker. It doesn't have all the sophistication of NEURON, but it's a lot easier to use. Has a robust and intuitive user interface. Doesn't scale especially well, and doesn't make full use of the GPU, but it's still a very nice environment, especially for small models. Nest is great when you're beginning a research effort and you just want to try a few things, you can get instant results on the screen in seconds. |
| Neutral | This is a Python layer around the NEURON simulator. It installs and works fine, but it gets a neutral on the basis that NEURON won't install on Windows. (It'll work if you debug the installation shell scripts, but no one should have to do that, it's Neuron's responsibility and they're releasing untested code which is a bad thing!) |
| Thumbs Down | Primitive and outdated. Bit of a toy. Complicated install, isn't being maintained. |
| Thumbs Up! | NeuroML isn't a simulator, it's a language for moving simulations and network definitions back and forth between different simulators. It's listed here because this activity is so important and so frequent in simulations. The ability to move data around is a must-have for any kind of work in computational neuroscience. One usually finds it's helpful to have a Python interpreter nearby, and Python has a rich ecosystem for parsing and editing formats like XML and JSON. Hopefully one never needs to actually read or edit these files, but when one does, a static intermediate representation is a blessing. |
| Neutral | An industry standard. Won't install on Windows, the install script is buggy. If you can get it to work (by manually editing shell scripts), it works nicely. The computational engine has been refined over 20 years, and it does things the others won't. However it doesn't handle fields correctly, it's not a finite element tool. It's a traditional simulator, and it excels in that area, and scales well. There are several tools available that can import and export simulations from other systems to Neuron. |
| Neutral | This is a great little educational tool that lets you look at the inside of an axon. It shows you ion channels and current flows. |
| Thumbs Down | Great little program, but isn't being maintained anymore. It's a stochastic simulator for ion channels, but it won't install or run on Windows. |
| Neutral | Simbrain is a lightweight simulator with a basic user interface. It doesn't import or export things, it's more designed for quick visualizations for the internet. Not really a professional tool, it's kind of halfway between educational and pre-publication. It has an interesting concept of "worlds" that has to do with the sensorium, for instance there is an Odor World and so on. There is an equivalent concept on the motor side, for speech and text. |
| Neutral | This program does non-meshed reaction-diffusion simulations in synapses and the areas around synapses. The latest GitHub is here. STEPS install from Python but has some major dependencies (like OpenBLAS, which is finicky). It does have a Docker version, available here. |
| Neutral | This is an interesting little tool that generates NEURON scripts from Trees morphologies. It hasn't been maintained in 8 years so it really should get a thumbs down, but the concept is so useful that it becomes a reference for anyone trying to automatically generate NEURON scripts. |
| Neutral | Virtual Brain is an expansive simulator from EBRAINS (the European consortium that also helped sponsor the Blue Brain project). It has dozens of dependencies (including Fortran and Kerberos). However it has some advanced functionality, like Slurm clustering. I'm not aware it does anything that others can't do (including interfacing with sensors, in case you want to get EEG or MEG data directly into the simulator), but it's definitely worth a checkout. It does some nice things with registration and it shows you brain maps that look a lot like fMRI, and it even has a volume conduction model. It gets a neutral on the basis of its enormous number of dependencies. |