A set of tools for fitting single- or multi-compartment neuron models parameters, using NSGA2 or Krayzman's dynamically weighted multi-objective optimization

Scripts, templates, and helpers to fit neuron model

├── README.md	This file
├── pyneuronautofit	module with extension of inspyred
│   ├── autofit.py	helper functions
│   ├── evaluator.py	evaluates a model
│   ├── fitter.py	main script for fitting
│   ├── __init__.py
│   ├── __main__.py -> fitter.py	: just a link for python -m pyneuronautofit
│   └── runandtest.py	: runs and tests a model can be call independently
├── scripts	: directory with useful scripts
│   ├── recovery-archive.sh	: recover archive if run was aborted
│   └── unique-in-ArXive.py	: collects only unique models
└── templates
. └── project.py	templay of a project file with all settings

Components

Evaluator

The Evaluator can perform analysis of the data and copare two data set against each other What kind of analysis it will perform is defined by mod variable. The mod variable is a string with one or more upper-case letters, each for specific analysis.

key	description
A	average spike shape during stimulus
C	distance between voltages during stimulus
D	distance between voltages during after stimulus tails
S	spike shapes during stimulus
T	spike times
R	resting potential
L	post-stimulus tail statistics
M	voltage stimulus statistics
N	number of spikes
O	Just total number of spikes
P	difference in probability dencity on v,dv/dt plane weighted by 1 - target_dencity/sum(target_dencity)
Q	the same as P but only during stimulus.
U	squared error of subthreshold voltage
V	distance between voltages
W	spike width during stimulus
Z	distance between voltages with zooming weight on spikes

$python -m pyneuronautofit -h
Usage: __main__.py [flags] input_file_with_currents_and_target_stats (abf,npz,or json)

Options:
  -h, --help            show this help message and exit

  Fitting:
    Parameters related to Evolutionary Optimization

    -A ALGOR, --algorithm=ALGOR
                        Algorithm for multiobjective evaluation. It can be:
                        Krayzman - for Krayzman's fitness weighting; NSGA2 -
                        for Pareto nondominate selection; Max - for max scaled
                        summation; PsitiveCor - positive correlation (the same
                        as Krayzman's procedure, but with goal of make all
                        correlations positive).Algorithm can be given by first
                        letter K, N, M or P correspondingly. (Default is K)
    -P PSZ, --population-size=PSZ
                        population size (default 256). If it is a negative
                        number: the population size is the length of the
                        fitness vector multiple by absolute value of this
                        option.
    -G NGN, --number-generation=NGN
                        number of generation (default 256)
    -E ELITES, --number-elites=ELITES
                        number of elites in the replacement (default 32)
    -L, --off-log-scale
                        enable log scaling
    -I INITPOP, --init-population=INITPOP
                        file with a set of initial population
    -N KRTHR, --Krayzman-threshould=KRTHR
                        Threshould for Krayzman's iteration procedure of
                        weights adaptation (default 0.05)
    -U UPDATE, --scales-update=UPDATE
                        vector length * this scale is number of fitness
                        vectors before update Krayzman's weights or max
                        scalers (default 10)
    -y, --norm-space    normalize space under the curve
    -H, --hold-weights-normalization
                        hold weights without normalization in iteration
                        procedure (default disable)
    -b BOUNDKGA, --bound-Krayzman-weights=BOUNDKGA
                        bound weights by [1/x,x] (default disable)
    -M MRATE, --mutation-rate=MRATE
                        Basic mutation rate (default 10%%)
    -S AMSLOPE, --adaptive-mutation-slope=AMSLOPE
                        Adaptive mutation slope
    -Q VPVSIZE, --v-dvdt-hist-size=VPVSIZE
                        v dv/dt histogram size (default 12)
    -J, --inJect-elits  enable dynamic elits

  Model:
    Conditions for model running and evaluation

    -m EMODE, --eval-mode=EMODE
                        mode for evaluation T-spike time, S-spike shape,
                        U-subthreshould voltage dynamics, W-spike width, R -
                        resting potential, L - post-stimulus tail, M - voltage
                        stimulus statistics, A - average spike shape, N -
                        number of spikes (default RAMN)
    -k EMASK, --eval-mask=EMASK
                        mask to limit analysis
    -c ESPC, --spike-count=ESPC
                        number of spikes for evaluation (2)
    -t ETHSH, --spike-threshold=ETHSH
                        spike threshold (default 0.)
    -l ELEFT, --left-spike-samples=ELEFT
                        left window of spike (default 70)
    -r ERGHT, --right-spike-samples=ERGHT
                        right window of spike (default 140)
    -q TEMP, --temperature=TEMP
                        temperature (default 35)
    -z SPWTGH, --spike-Zoom=SPWTGH
                        if positive absolute weight of voltage diff during
                        spike; if negative relataed scaler
    -e, --collapse-diff
                        Collapse difference between a model and data in a
                        vector with size = number of tests (i.e. for  -m RAMNT
                        the diff vector will be length 5)

  Run:
    Options for entire EC running and logging

    -n NTH, --number-threads=NTH
                        number of threads (default None - autodetection)
    --dt=SIMDT          if positive absolute simulation dt; if negative scaler
                        for recorded dt
    -v LL, --log-level=LL
                        Level of logging.[CRITICAL, ERROR, WARNING, INFO, or
                        DEBUG] (default INFO)
    -u, --log-to-screen
                        log to screen
    -Z, --Krayzman-debug
                        enable debug dump for adaptation weight
    -p NCH, --printed-checkpoints=NCH
                        print out checkpoints every # generation (do not print
                        out if negative)
    -d DCH, --dump-checkpoints=DCH
                        dump out checkpoints into checkpoint file every #
                        generation (do not dump out if negative, default 8)
    -i RITER, --iteration=RITER
                        adds iteration number to the runs stamp
    -a RSTEMP, --run-stamp=RSTEMP
                        Use this run stamp instead of generated
    --slurm-id=SLURMID  Add SLURM ID into timestamp
    --log-population    record population into log file
    --log-archive       record archive into log file
    --dry-run           exit after init everything