Definition of "pre" property of :Neuron

49 views
Skip to first unread message

Ryosuke Tanaka

unread,
Feb 16, 2022, 6:52:51 PM2/16/22
to neuPrint
Hi all,

I just came to realize that "pre" property of a :Neuron can be much smaller than the sum of the its connection weights (in the hemibrain v1.2.1 dataset).

For example, the following query

MATCH (a:Neuron)-[w:ConnectsTo]->(b:Neuron)
WHERE a.bodyId = 5813047992
RETURN a.pre, sum(w.weight)

returns respectively 346 and 868. 
Even if I add b.type IS NOT NULL to the condition, I still get sum(w.weight) = 658.
Based on the description in the user guide, I assumed these numbers to be equal. Does anyone happen to know what exactly "pre" is counting? I am sorry if I am missing something basic.

Best regards,
Ryosuke


Stuart Berg

unread,
Feb 17, 2022, 12:07:58 AM2/17/22
to neuPrint
Hi Ryosuke

I just came to realize that "pre" property of a :Neuron can be much smaller than the sum of the its connection weights (in the hemibrain v1.2.1 dataset).

In drosophila, each presynaptic site (t-bar) is typically connected to multiple postsynaptic sites.  The 'pre' property tells you how many t-bars are present in the neuron, but not how many downstream partners all of its t-bars have in total.  For that, you can use the 'downstream' property.

Here's a version of your query where I've made two modifications:
1. I match all downstream :Segment objects, rather than only :Neuron objects.  In neuprint, all :Neurons are :Segments, but most :Segments are not large enough to qualify as a :Neuron.  The 'downstream' property refers to all postsynaptic sites from the neuron's own t-bars, regardless of whether those sites belong to :Neurons or merely :Segments.
2. I added the 'downstream' property to the results.  As you can see, it matches the sum(w.weight) result.

    MATCH (a:Neuron)-[w:ConnectsTo]->(b:Segment)
    WHERE a.bodyId = 5813047992
    RETURN a.pre, a.downstream, sum(w.weight)

That results in the following values:

a.pre           346
a.downstream   1861
sum(w.weight)  1861

BTW, if it's helpful to visualize a t-bar and its postsynaptic sites, here's an example t-bar from the neuron in your query:

Best regards,
Stuart


Reply all
Reply to author
Forward
0 new messages