Probe operators¶

Probe operators map simulation state into named readouts.

Emitter -> Source -> Field -> Probe -> Objective -> Optimizer

Operators¶

Operator	Output	Typical shape
SPK	spike matrix or events	`[time, neuron]`
Vm	voltage/state trace	`[time, neuron]`
Source	source/current proxy	`[time, source]`
LFP-proxy	laminar field proxy	`[time, contact]`
CSD-proxy	spatial source/divergence proxy	`[time, contact]`
EEG-proxy	linear channel projection	`[time, channel]`
MEG-proxy	linear magnetic-channel projection	`[time, channel]`
EMM-proxy	normalized activity-cost proxy	`[time]`

Mathematical Forms¶

Computational scaffolds define the following proxy readout operations:

SPK: $y_{spk}[t, k] = \sum_i \delta(t - t_{i,k})$
Vm: $y_{vm}[t, k] = V_k[t]$
Source: $y_{src}[t, k] = J_k[t]$
LFP-proxy: $y_{lfp}[t, k] = \sum_i \alpha_{ik} V_i[t]$
CSD-proxy: $y_{csd}[t, k] = \nabla^2 y_{lfp}[t, k]$
EEG-proxy: $y_{eeg}[t, k] = \sum_i \beta_{ik} y_{lfp}[t, i]$
MEG-proxy: $y_{meg}[t, k] = \sum_i \gamma_{ik} y_{src}[t, i]$
EMM-proxy: $y_{emm}[t] = \frac{1}{N} \sum_k |y_{src}[t, k]|

name: string
kind: spk | vm | source | lfp_proxy | csd_proxy | eeg_proxy | meg_proxy | emm_proxy
method: string
data_shape: list[int]
units_or_status: string
operator_status: simulated_proxy
calibration_status: string

Example¶

model = jtfne.construct(cfg)
signals = jtfne.simulate(model, duration_ms=1000.0, dt_ms=0.1, seed=7)
readout = model.probe(signals)
print(readout.to_dict())