Statistics

This notebook shows show the do some basic statistics on Neuropixels ChannelMap and its blueprint.

Number of selected electrodes

For a channelmap

[5]:

from neurocarto.probe_npx import ChannelMap

chmap = ChannelMap(24)
len(chmap) # empty map

[5]:

[10]:

chmap.add_electrode((0,0,0))
len(chmap)

[10]:

For a blueprint

[11]:

from neurocarto.probe_npx import NpxProbeDesp

D = NpxProbeDesp()

# all electrodes, chmap here used as a probe type reference, we do not use its content inside.
blueprint = D.all_electrodes(chmap)
len([e for e in blueprint if e.state == D.STATE_USED])

[11]:

[13]:

# Get selected electrodes
# This function will read the content of chmap
len(D.all_channels(chmap))

[13]:

Number of electrode set with full-density category

[16]:

blueprint = D.load_blueprint('Fig3_example.blueprint.npy', chmap)

len([e for e in blueprint if e.category == D.CATE_FULL])

[16]:

Area Efficiency

we define the area efficiency \(Aeff\) of a given blueprint \(Q\) and its outcomes channel map \(S\) as

\(Aeff(S|Q) = \cfrac{selected(S|Q)}{request(Q)}\), \(0\) if \(request(Q)=0\)

Where \(selected\) and \(request\) are defined below:

[18]:

def selected(S, Q):
    """How many electrodes are selected in density-category."""
    # we do not use S here, because Q contains the selecting information.
    v1 = len([e for e in Q if e.state == D.STATE_USED and e.category == D.CATE_SET])
    v2 = len([e for e in Q if e.state == D.STATE_USED and e.category == D.CATE_FULL])
    v3 = len([e for e in Q if e.state == D.STATE_USED and e.category == D.CATE_HALF])
    v4 = len([e for e in Q if e.state == D.STATE_USED and e.category == D.CATE_QUARTER])
    return v1 + v2 + v3 + v4

def request(Q):
    """How many electrodes are requested in density-category."""
    v1 = len([e for e in Q if e.category == D.CATE_SET])
    v2 = len([e for e in Q if e.category == D.CATE_FULL])
    v3 = len([e for e in Q if e.category == D.CATE_HALF])
    v4 = len([e for e in Q if e.category == D.CATE_QUARTER])
    return v1 + v2 + v3/2 + v4/4

When \(Aeff < 1\) , it means area are not well used on average. When \(Aeff > 1\) , it means area are over-used on average.

Channel Efficiency

We define the channel efficiency \(Ceff\) of a given blueprint \(Q\) and its outcomes channel map \(S\) as

\(Ceff(S|Q) = min \{ Aeff(S|Q), \cfrac{1}{Aeff(S|Q)} \}\), \(0\) if \(Aeff(S|Q) = 0\)

We define the channel efficiency \(Ceff\) of a blueprint \(Q\) as

\(Ceff(Q) = max \{ Ceff(S|Q) | S \in \mathcal{S} \}\)

where \(\mathcal{S}\) is all possible outcomes from the given blueprint \(Q\).