Ashwin Answers Mystic Questions
@KrzysztofKruk asks:
I wonder, what are those seemingly empty quasi-spherical structures, which we can see in some of the synapses. In some cells there are barely any of those and in others almost every synapse has one. For example, this one: 
Ashwin:
My guess is that those structures are in fact mitochondria that for some reason were not well stained. It fit in line with the fact that typically we tend to see the mitochondria close to a synapse.
@Atani Asks:
What is msty’s overall accuracy?
Ashwin:
I think in the early days, we had determined its accuracy on average to be ~70-80%. It varies depending on whether the segment being traced is an axon or a dendrite. In general, axons tend to be much more difficult as compared to dendrites. This is also related to how much training went into the neural nets, more is better of course.
@susi asks:
How many types of neurons we have traced so far? Do they have typical locations? Special functions? Can you give us typical overall shapes of these?
Ashwin:
You can check out information on the number and morphologies of the neurons here.
@susi asks:
Twister once gave me a graph, does this represent the majority of cells we trace? (Dendrite part normally just out of dataset ?)
If I understand the image correctly, I think twist was highlighting the dendrites (green) and axons (blue/red). In general this tends to be the morphology of most neurons, short stubby dendrites and long thin axons. They also typically tend to run perpendicular to each other.
@susi asks:
Are we still tracing neurons which communicate with integrator-neurons?
Absolutely! This is what is amazing! All that is being traced are in someway taking to the original 22.
@susi asks:
Could you again explain ipsilateral and contralateral (if this is of importance in context of the cells we actually trace)?
Ipsilateral → means on the same side.
Contralateral → mean on the opposite side.
Since we are tracing neurons only from one side of the fish, if we refer to a neuron as an ipsilaterally projecting neurons, that mean it projects its axons to within the same side that we are reconstructing. If we call it a contralateral projecting neuron, it either sends it axon to the opposite side, it originates from the opposite side.
@ggreminder asks:
What is the big picture of what we have mapped so far? Why hindbrain?
The big picture is that we are interested in studying the neural circuits that underlie eye movements. All eye movements are controlled by neurons in the hindbrain (brain stem). By reconstructing these neurons, we are able to precisely tell which neurons input to these eye-movement related neurons, who are the outputs of the eye-movement related neurons. In effect, we can trace the exact path of information flow that happens when we move our eyes.