Braincoder documentation¶

braincoder provides tools to build encoding models, convolve them with hemodynamic response functions, and decode stimuli from neural data.

Quick start¶

The example below mirrors the steps you will take in most analyses: prepare a paradigm, instantiate a model, simulate data, and fit parameters.

import numpy as np
import pandas as pd
from braincoder.models import LinearModelWithBaseline
from braincoder.optimize import ParameterFitter

n_tp = 200
paradigm = pd.DataFrame(
    {"stimulus": np.sin(np.linspace(0, 4 * np.pi, n_tp))},
    index=pd.Index(np.arange(n_tp), name="time"),
)

true_params = pd.DataFrame(
    {"baseline": [0.0, 0.5]},
    index=pd.Index(["voxel_1", "voxel_2"], name="voxel"),
)

model = LinearModelWithBaseline(paradigm=paradigm, parameters=true_params)
simulated = model.simulate(
    paradigm=paradigm,
    parameters=true_params,
    noise=0.1,
)

fitter = ParameterFitter(model=model, data=simulated, paradigm=paradigm)
estimates = fitter.fit(max_n_iterations=200, learning_rate=0.05)

print(estimates)

More resources¶

What’s new — release notes; 0.6.0 adds calibrated decoding (Noise models for decoding: low-rank Ω and temperature).
Concepts & Architecture — architectural overview and workflow tips.
Noise models for decoding: low-rank Ω and temperature — calibrated decoding: low-rank Ω with cross-validated rank and temperature.
API Reference — autodoc for the most used modules.
Tutorial — hands-on lessons that walk through realistic encoding/decode pipelines.
Examples — gallery of runnable scripts.