pyhrf.jde.nrl.bigaussian module¶
-
class
pyhrf.jde.nrl.bigaussian.
BiGaussMixtureParamsSampler
(do_sampling=True, use_true_value=False, val_ini=None, hyper_prior_type='Jeffreys', activ_thresh=4.0, var_ci_pr_alpha=2.04, var_ci_pr_beta=0.5, var_ca_pr_alpha=2.01, var_ca_pr_beta=0.5, mean_ca_pr_mean=5.0, mean_ca_pr_var=20.0)¶ Bases:
pyhrf.xmlio.Initable
,pyhrf.jde.samplerbase.GibbsSamplerVariable
#TODO : comment
-
I_MEAN_CA
= 0¶
-
I_VAR_CA
= 1¶
-
I_VAR_CI
= 2¶
-
L_CA
= 1¶
-
L_CI
= 0¶
-
NB_PARAMS
= 3¶
-
PARAMS_NAMES
= ['Mean_Activ', 'Var_Activ', 'Var_Inactiv']¶
-
checkAndSetInitValue
(variables)¶
-
computeWithJeffreyPriors
(j, cardCIj, cardCAj)¶
-
computeWithProperPriors
(j, cardCIj, cardCAj)¶
-
finalizeSampling
()¶
-
getCurrentMeans
()¶
-
getCurrentVars
()¶
-
getOutputs
()¶
-
get_string_value
(v)¶
-
linkToData
(dataInput)¶
-
parametersComments
= {'activ_thresh': 'Threshold for the max activ mean above which the region is considered activating', 'hyper_prior_type': "Either 'proper' or 'Jeffreys'"}¶
-
parametersToShow
= []¶
-
sampleNextInternal
(variables)¶
-
updateObsersables
()¶
-
-
class
pyhrf.jde.nrl.bigaussian.
BiGaussMixtureParamsSamplerWithRelVar
(do_sampling=True, use_true_value=False, val_ini=None, hyper_prior_type='Jeffreys', activ_thresh=4.0, var_ci_pr_alpha=2.04, var_ci_pr_beta=0.5, var_ca_pr_alpha=2.01, var_ca_pr_beta=0.5, mean_ca_pr_mean=5.0, mean_ca_pr_var=20.0)¶ Bases:
pyhrf.jde.nrl.bigaussian.BiGaussMixtureParamsSampler
-
computeWithProperPriorsWithRelVar
(nrlsj, j, cardCIj, cardCAj, wj)¶
-
sampleNextInternal
(variables)¶
-
-
class
pyhrf.jde.nrl.bigaussian.
BiGaussMixtureParamsSamplerWithRelVar_OLD
(do_sampling=True, use_true_value=False, val_ini=None, hyper_prior_type='Jeffreys', activ_thresh=4.0, var_ci_pr_alpha=2.04, var_ci_pr_beta=0.5, var_ca_pr_alpha=2.01, var_ca_pr_beta=0.5, mean_ca_pr_mean=5.0, mean_ca_pr_var=20.0)¶ Bases:
pyhrf.jde.nrl.bigaussian.BiGaussMixtureParamsSampler
-
computeWithProperPriorsWithRelVar
(nrlsj, j, cardCIj, cardCAj, wj)¶
-
sampleNextInternal
(variables)¶
-
-
class
pyhrf.jde.nrl.bigaussian.
MixtureWeightsSampler
(do_sampling=True, use_true_value=False, val_ini=None)¶ Bases:
pyhrf.xmlio.Initable
,pyhrf.jde.samplerbase.GibbsSamplerVariable
#TODO : comment
-
checkAndSetInitValue
(variables)¶
-
getOutputs
()¶
-
linkToData
(dataInput)¶
-
sampleNextInternal
(variables)¶
-
-
class
pyhrf.jde.nrl.bigaussian.
NRLSampler
(do_sampling=True, val_ini=None, contrasts={}, do_label_sampling=True, use_true_nrls=False, use_true_labels=False, labels_ini=None, ppm_proba_threshold=0.05, ppm_value_threshold=0, ppm_value_multi_threshold=array([ 0., 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1., 1.1, 1.2, 1.3, 1.4, 1.5, 1.6, 1.7, 1.8, 1.9, 2., 2.1, 2.2, 2.3, 2.4, 2.5, 2.6, 2.7, 2.8, 2.9, 3., 3.1, 3.2, 3.3, 3.4, 3.5, 3.6, 3.7, 3.8, 3.9, 4. ]), mean_activation_threshold=4, rescale_results=False, wip_variance_computation=False)¶ Bases:
pyhrf.xmlio.Initable
,pyhrf.jde.samplerbase.GibbsSamplerVariable
Class handling the Gibbs sampling of Neural Response Levels with a prior bi-gaussian mixture model. It handles independent and spatial versions. Refs : Vincent 2010 IEEE TMI, Makni 2008 Neuroimage, Sockel 2009 ICASSP #TODO : comment attributes
-
CLASSES
= array([0, 1])¶
-
CLASS_NAMES
= ['inactiv', 'activ']¶
-
FALSE_NEG
= 3¶
-
FALSE_POS
= 2¶
-
L_CA
= 1¶
-
L_CI
= 0¶
-
PPMcalculus
(threshold_value, apost_mean_activ, apost_var_activ, apost_mean_inactiv, apost_var_inactiv, labels_activ, labels_inactiv)¶ Function to calculate the probability that the nrl in voxel j, condition m, is superior to a given hreshold_value
-
ThresholdPPM
(proba_voxel, threshold_pval)¶
-
calcFracLambdaTilde
(cond, c1, c2, variables)¶
-
checkAndSetInitLabels
(variables)¶
-
checkAndSetInitNRL
(variables)¶
-
checkAndSetInitValue
(variables)¶
-
cleanMemory
()¶
-
cleanObservables
()¶
-
computeAA
(nrls, destaa)¶
-
computeComponentsApost
(variables, j, gTQg)¶
-
computeContrasts
()¶
-
computeVarXhtQ
(h, varXQ)¶
-
computeVarYTildeOpt
(varXh)¶
-
compute_summary_stats
()¶
-
countLabels
(labels, voxIdx, cardClass)¶
-
finalizeSampling
()¶
-
getClassifRate
()¶
-
getFinalLabels
(thres=None)¶
-
getOutputs
()¶
-
getRocData
(dthres=0.005)¶
-
get_final_summary
()¶
-
initObservables
()¶
-
init_contrasts
()¶
-
linkToData
(dataInput)¶
-
markWrongLabels
(labels)¶
-
parametersComments
= {'contrasts': 'Define contrasts as arithmetic expressions.\nCondition names used in expressions must be consistent with those specified in session data above'}¶
-
parametersToShow
= ['contrasts']¶
-
printState
(_)¶
-
reportDetection
()¶
-
sampleLabels
(cond, variables)¶
-
sampleNextAlt
(variables)¶
-
sampleNextInternal
(variables)¶
-
sampleNrlsParallel
(varXh, rb, h, varLambda, varCI, varCA, meanCA, gTQg, variables)¶
-
sampleNrlsSerial
(rb, h, varCI, varCA, meanCA, gTQg, variables)¶
-
samplingWarmUp
(variables)¶ #TODO : comment
-
saveCurrentValue
(it)¶
-
saveObservables
(it)¶
-
updateObsersables
()¶
-
-
class
pyhrf.jde.nrl.bigaussian.
NRLSamplerWithRelVar
(do_sampling=True, val_ini=None, contrasts={}, do_label_sampling=True, use_true_nrls=False, use_true_labels=False, labels_ini=None, ppm_proba_threshold=0.05, ppm_value_threshold=0, ppm_value_multi_threshold=array([ 0., 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1., 1.1, 1.2, 1.3, 1.4, 1.5, 1.6, 1.7, 1.8, 1.9, 2., 2.1, 2.2, 2.3, 2.4, 2.5, 2.6, 2.7, 2.8, 2.9, 3., 3.1, 3.2, 3.3, 3.4, 3.5, 3.6, 3.7, 3.8, 3.9, 4. ]), mean_activation_threshold=4, rescale_results=False, wip_variance_computation=False)¶ Bases:
pyhrf.jde.nrl.bigaussian.NRLSampler
-
calcFracLambdaTildeWithIRRelCond
(cond, c1, c2, variables, nbVox, moyqvoxj, t1, t2)¶
-
calcFracLambdaTildeWithRelCond
(l, nbVox, moyqvoxj, t1, t2)¶
-
computeComponentsApostWithRelVar
(variables, j, gTQg, w)¶
-
computeSumWAxh
(wa, varXh)¶
-
computeVarYTildeOptWithRelVar
(varXh, w)¶
-
computeWA
(a, w, wa)¶
-
computemoyqvox
(cardClass, nbVox)¶ Compute mean of labels in ROI (without the label of voxel i)
-
createWAxh
(aXh, w)¶
-
deltaWCorr0
(nbVox, moyqvoxj, t1, t2)¶
-
deltaWCorr1
(nbVox, moyqvoxj, t1, t2)¶
-
sampleLabelsWithRelVar
(cond, variables)¶
-
sampleNextInternal
(variables)¶
-
sampleNrlsParallelWithRelVar
(varXh, rb, h, varLambda, varCI, varCA, meanCA, gTQg, variables, w)¶
-
sampleNrlsSerialWithRelVar
(rb, h, gTQg, variables, w, t1, t2)¶
-
samplingWarmUp
(variables)¶
-
subtractYtildeWithRelVar
()¶
-
-
class
pyhrf.jde.nrl.bigaussian.
NRL_Multi_Sess_Sampler
(parameters=None, xmlHandler=None, xmlLabel=None, xmlComment=None)¶ Bases:
pyhrf.jde.samplerbase.GibbsSamplerVariable
-
P_OUTPUT_NRL
= 'writeResponsesOutput'¶
-
P_SAMPLE_FLAG
= 'sampleFlag'¶
-
P_TrueNrlFilename
= 'TrueNrlFilename'¶
-
P_USE_TRUE_NRLS
= 'useTrueNrls'¶
-
P_VAL_INI
= 'initialValue'¶
-
checkAndSetInitValue
(variables)¶
-
cleanMemory
()¶
-
computeAA
(nrls, destaa)¶
-
computeComponentsApost
(variables, m, varXh, s)¶
-
computeVarYTildeSessionOpt
(varXh, s)¶
-
defaultParameters
= {'writeResponsesOutput': True, 'initialValue': None, 'TrueNrlFilename': None, 'useTrueNrls': False, 'sampleFlag': True}¶
-
finalizeSampling
()¶
-
getOutputs
()¶
-
linkToData
(dataInput)¶
-
parametersComments
= {'TrueNrlFilename': 'Define the filename of simulated NRLs.\nIt is taken into account when NRLs is not sampled.'}¶
-
parametersToShow
= ['writeResponsesOutput']¶
-
sampleNextAlt
(variables)¶
-
sampleNextInternal
(variables)¶
-
samplingWarmUp
(variables)¶ #TODO : comment
-
saveCurrentValue
(it)¶
-
-
class
pyhrf.jde.nrl.bigaussian.
Variance_GaussianNRL_Multi_Sess
(parameters=None, xmlHandler=None, xmlLabel=None, xmlComment=None)¶ Bases:
pyhrf.jde.samplerbase.GibbsSamplerVariable
-
P_SAMPLE_FLAG
= 'sampleFlag'¶
-
P_USE_TRUE_VALUE
= 'useTrueValue'¶
-
P_VAL_INI
= 'initialValue'¶
-
checkAndSetInitValue
(variables)¶
-
defaultParameters
= {'useTrueValue': False, 'initialValue': array([ 1.]), 'sampleFlag': False}¶
-
linkToData
(dataInput)¶
-
parametersToShow
= ['useTrueValue']¶
-
sampleNextInternal
(variables)¶
-