# Dombrovski*, Luna, Hallquist*, Nature Communications, 2020
# brain-to-behavior analyses with anterior (low entropy) hippocampal cluster betas
# first run beta_cluster_import_pca_clean.R if not run once already
library(tidyverse)
## ── Attaching packages ─────────────────────────────────────── tidyverse 1.3.0 ──
## ✓ ggplot2 3.3.3 ✓ purrr 0.3.4
## ✓ tibble 3.0.5 ✓ dplyr 1.0.3
## ✓ tidyr 1.1.2 ✓ stringr 1.4.0
## ✓ readr 1.4.0 ✓ forcats 0.5.0
## ── Conflicts ────────────────────────────────────────── tidyverse_conflicts() ──
## x dplyr::filter() masks stats::filter()
## x dplyr::lag() masks stats::lag()
library(psych)
##
## Attaching package: 'psych'
## The following objects are masked from 'package:ggplot2':
##
## %+%, alpha
library(lme4)
## Loading required package: Matrix
##
## Attaching package: 'Matrix'
## The following objects are masked from 'package:tidyr':
##
## expand, pack, unpack
library(lmerTest)
##
## Attaching package: 'lmerTest'
## The following object is masked from 'package:lme4':
##
## lmer
## The following object is masked from 'package:stats':
##
## step
library(car)
## Loading required package: carData
## Registered S3 methods overwritten by 'car':
## method from
## influence.merMod lme4
## cooks.distance.influence.merMod lme4
## dfbeta.influence.merMod lme4
## dfbetas.influence.merMod lme4
##
## Attaching package: 'car'
## The following object is masked from 'package:psych':
##
## logit
## The following object is masked from 'package:dplyr':
##
## recode
## The following object is masked from 'package:purrr':
##
## some
library(emmeans)
# clock_folder <- "~/Data_Analysis/clock_analysis" #michael
clock_folder <- "~/code/clock_analysis_mlm" #alex
setwd(file.path(clock_folder, 'fmri/keuka_brain_behavior_analyses/'))
### load data
# load('trial_df_and_vhdkfpe_clusters.Rdata')
# cleaner version with only H, PE and uncertainty trial vars
unsmoothed = F
if (unsmoothed) {
load('trial_df_and_vh_pe_clusters_u_unsmoothed.Rdata')
} else { load('trial_df_and_vh_pe_clusters_u.Rdata') }
# inspect behavioral data
# make sure no one always responds immediately
# raw, all subjects and trials
ggplot(df, aes(trial, rt_csv, color = rewFunc)) + geom_line() + facet_wrap(~id)
# smoothed
ggplot(df, aes(run_trial, rt_csv, color = rewFunc)) + geom_smooth(method = "gam") + facet_wrap(~id)
## `geom_smooth()` using formula 'y ~ s(x, bs = "cs")'
# inspect AH betas
ggplot(df, aes(h_HippAntL_neg)) + geom_histogram() + xlab("Anterior hippocampal response to low entropy")
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
# how do we model RT swings? ----
# naive definition: abs(RT - RT_lag)
hist(df$rt_swing)
# ugly, 0-inflated distribution
# We will model RT swings as the [negative] slope of RT_lag, RT ~ RT_lag
## Simple model with condition ----
# trial_neg_inv_sc: scale(-1/trial) to better describe the learning curve
model_condition <- lmer(rt_csv_sc ~ trial_neg_inv_sc + rt_lag_sc + rewFunc + last_outcome +
(1|id/run), df %>% filter(!is.na(rt_vmax_lag_sc)))
summary(model_condition)
## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: rt_csv_sc ~ trial_neg_inv_sc + rt_lag_sc + rewFunc + last_outcome +
## (1 | id/run)
## Data: df %>% filter(!is.na(rt_vmax_lag_sc))
##
## REML criterion at convergence: 63227.7
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.0793 -0.5752 -0.0668 0.5492 4.4419
##
## Random effects:
## Groups Name Variance Std.Dev.
## run:id (Intercept) 0.06851 0.2618
## id (Intercept) 0.06866 0.2620
## Residual 0.56890 0.7543
## Number of obs: 27253, groups: run:id, 560; id, 70
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 5.537e-02 4.627e-02 2.181e+02 1.197 0.2327
## trial_neg_inv_sc 4.420e-02 8.618e-03 2.671e+04 5.128 2.94e-07 ***
## rt_lag_sc 3.342e-01 5.862e-03 2.695e+04 57.008 < 2e-16 ***
## rewFuncCEVR 1.229e-01 4.803e-02 4.612e+02 2.558 0.0108 *
## rewFuncDEV -2.022e-01 3.913e-02 4.572e+02 -5.168 3.54e-07 ***
## rewFuncIEV 2.182e-01 3.915e-02 4.581e+02 5.575 4.24e-08 ***
## last_outcomeOmission -1.835e-01 1.038e-02 2.687e+04 -17.680 < 2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) trl___ rt_lg_ rFCEVR rwFDEV rwFIEV
## trl_ng_nv_s -0.021
## rt_lag_sc 0.024 -0.104
## rewFuncCEVR -0.509 0.003 0.009
## rewFuncDEV -0.630 -0.008 0.049 0.611
## rewFuncIEV -0.635 0.003 -0.057 0.606 0.744
## lst_tcmOmss -0.081 -0.017 -0.233 -0.080 -0.022 0.038
## With RT_Vmax ----
# RT_Vmax is the best response time according to the SCEPTIC model
# its effect reflects exploitation (convergence on the best)
model_RT_Vmax <- lmer(rt_csv_sc ~ trial_neg_inv_sc + rt_lag_sc + rt_vmax_lag_sc + last_outcome +
(1|id/run), df %>% filter(!is.na(rt_vmax_lag_sc)))
summary(model_RT_Vmax)
## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula:
## rt_csv_sc ~ trial_neg_inv_sc + rt_lag_sc + rt_vmax_lag_sc + last_outcome +
## (1 | id/run)
## Data: df %>% filter(!is.na(rt_vmax_lag_sc))
##
## REML criterion at convergence: 62993.7
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.9792 -0.5687 -0.0687 0.5441 4.4313
##
## Random effects:
## Groups Name Variance Std.Dev.
## run:id (Intercept) 0.06970 0.2640
## id (Intercept) 0.04433 0.2106
## Residual 0.56443 0.7513
## Number of obs: 27253, groups: run:id, 560; id, 70
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 8.111e-02 2.820e-02 7.017e+01 2.876 0.00533 **
## trial_neg_inv_sc 4.084e-02 8.583e-03 2.665e+04 4.759 1.96e-06 ***
## rt_lag_sc 3.024e-01 6.138e-03 2.721e+04 49.268 < 2e-16 ***
## rt_vmax_lag_sc 1.458e-01 6.752e-03 2.373e+04 21.599 < 2e-16 ***
## last_outcomeOmission -1.931e-01 1.023e-02 2.713e+04 -18.868 < 2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) trl___ rt_lg_ rt_v__
## trl_ng_nv_s -0.037
## rt_lag_sc 0.033 -0.095
## rt_vmx_lg_s 0.001 -0.005 -0.340
## lst_tcmOmss -0.140 -0.018 -0.209 -0.003
# naturally, including RT V-max instead of condition improves the fit by 300 AIC points (not a true model comparison, not nested)
anova(model_RT_Vmax, model_condition)
## refitting model(s) with ML (instead of REML)
## Data: df %>% filter(!is.na(rt_vmax_lag_sc))
## Models:
## model_RT_Vmax: rt_csv_sc ~ trial_neg_inv_sc + rt_lag_sc + rt_vmax_lag_sc + last_outcome +
## model_RT_Vmax: (1 | id/run)
## model_condition: rt_csv_sc ~ trial_neg_inv_sc + rt_lag_sc + rewFunc + last_outcome +
## model_condition: (1 | id/run)
## npar AIC BIC logLik deviance Chisq Df Pr(>Chisq)
## model_RT_Vmax 8 62973 63038 -31478 62957
## model_condition 10 63204 63286 -31592 63184 0 2 1
## add 2-way interactions ----
model_RT_Vmax_interactions <- lmer(rt_csv_sc ~ (trial_neg_inv_sc + rt_lag_sc + rt_vmax_lag_sc + last_outcome)^2 +
(1|id/run), df %>% filter(!is.na(rt_vmax_lag_sc)))
summary(model_RT_Vmax_interactions)
## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: rt_csv_sc ~ (trial_neg_inv_sc + rt_lag_sc + rt_vmax_lag_sc +
## last_outcome)^2 + (1 | id/run)
## Data: df %>% filter(!is.na(rt_vmax_lag_sc))
##
## REML criterion at convergence: 61697.6
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.4550 -0.5502 -0.0760 0.5119 4.5605
##
## Random effects:
## Groups Name Variance Std.Dev.
## run:id (Intercept) 0.05764 0.2401
## id (Intercept) 0.04027 0.2007
## Residual 0.53871 0.7340
## Number of obs: 27253, groups: run:id, 560; id, 70
##
## Fixed effects:
## Estimate Std. Error df t value
## (Intercept) 1.077e-01 2.689e-02 7.205e+01 4.004
## trial_neg_inv_sc 9.928e-03 1.149e-02 2.668e+04 0.864
## rt_lag_sc 4.795e-01 8.039e-03 2.722e+04 59.646
## rt_vmax_lag_sc 1.240e-01 8.067e-03 2.564e+04 15.368
## last_outcomeOmission -1.545e-01 1.037e-02 2.703e+04 -14.898
## trial_neg_inv_sc:rt_lag_sc -6.981e-02 1.032e-02 2.676e+04 -6.762
## trial_neg_inv_sc:rt_vmax_lag_sc 7.185e-02 9.176e-03 2.718e+04 7.831
## trial_neg_inv_sc:last_outcomeOmission 1.203e-02 1.823e-02 2.668e+04 0.660
## rt_lag_sc:rt_vmax_lag_sc 1.907e-03 4.941e-03 2.668e+04 0.386
## rt_lag_sc:last_outcomeOmission -3.760e-01 1.150e-02 2.706e+04 -32.688
## rt_vmax_lag_sc:last_outcomeOmission 7.090e-02 1.166e-02 2.703e+04 6.083
## Pr(>|t|)
## (Intercept) 0.000149 ***
## trial_neg_inv_sc 0.387494
## rt_lag_sc < 2e-16 ***
## rt_vmax_lag_sc < 2e-16 ***
## last_outcomeOmission < 2e-16 ***
## trial_neg_inv_sc:rt_lag_sc 1.39e-11 ***
## trial_neg_inv_sc:rt_vmax_lag_sc 5.03e-15 ***
## trial_neg_inv_sc:last_outcomeOmission 0.509447
## rt_lag_sc:rt_vmax_lag_sc 0.699483
## rt_lag_sc:last_outcomeOmission < 2e-16 ***
## rt_vmax_lag_sc:last_outcomeOmission 1.20e-09 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) trl___ rt_lg_ rt_v__ lst_tO tr___:__ t___:___ t___:_O r__:__
## trl_ng_nv_s -0.053
## rt_lag_sc 0.059 -0.137
## rt_vmx_lg_s -0.003 -0.044 -0.391
## lst_tcmOmss -0.144 0.131 -0.108 -0.010
## trl_ng__:__ -0.037 0.420 -0.136 -0.060 0.032
## trl_n__:___ 0.013 -0.005 -0.002 -0.018 0.021 -0.260
## trl_ng__:_O 0.036 -0.604 0.084 0.027 -0.247 -0.187 -0.059
## rt_lg_s:___ -0.099 -0.061 -0.140 0.047 -0.003 0.007 -0.130 0.003
## rt_lg_sc:_O -0.030 0.039 -0.642 0.310 -0.074 -0.014 0.004 -0.077 0.071
## rt_vmx__:_O 0.027 0.005 0.361 -0.568 0.000 0.008 -0.011 0.029 -0.233
## r__:_O
## trl_ng_nv_s
## rt_lag_sc
## rt_vmx_lg_s
## lst_tcmOmss
## trl_ng__:__
## trl_n__:___
## trl_ng__:_O
## rt_lg_s:___
## rt_lg_sc:_O
## rt_vmx__:_O -0.555
# the interaction that helps the most is rt_lag * last_outcome
# win-stay/lose-shift
anova(model_RT_Vmax, model_RT_Vmax_interactions)
## refitting model(s) with ML (instead of REML)
## Data: df %>% filter(!is.na(rt_vmax_lag_sc))
## Models:
## model_RT_Vmax: rt_csv_sc ~ trial_neg_inv_sc + rt_lag_sc + rt_vmax_lag_sc + last_outcome +
## model_RT_Vmax: (1 | id/run)
## model_RT_Vmax_interactions: rt_csv_sc ~ (trial_neg_inv_sc + rt_lag_sc + rt_vmax_lag_sc +
## model_RT_Vmax_interactions: last_outcome)^2 + (1 | id/run)
## npar AIC BIC logLik deviance Chisq Df
## model_RT_Vmax 8 62973 63038 -31478 62957
## model_RT_Vmax_interactions 14 61644 61759 -30808 61616 1340.9 6
## Pr(>Chisq)
## model_RT_Vmax
## model_RT_Vmax_interactions < 2.2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
###############
# Main analyses including model-derived behavioral variables
# hippocampal model-based analysis
model_AH_beta_fMRI <- lmer(rt_csv_sc ~ (trial_neg_inv_sc + rt_lag_sc + rt_vmax_lag_sc + last_outcome +
v_max_wi_lag + v_entropy_wi + h_HippAntL_neg)^2 +
rt_lag_sc:last_outcome:h_HippAntL_neg +
rt_vmax_lag_sc:trial_neg_inv_sc:h_HippAntL_neg +
(1|id/run), df)
summary(model_AH_beta_fMRI)
## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: rt_csv_sc ~ (trial_neg_inv_sc + rt_lag_sc + rt_vmax_lag_sc +
## last_outcome + v_max_wi_lag + v_entropy_wi + h_HippAntL_neg)^2 +
## rt_lag_sc:last_outcome:h_HippAntL_neg + rt_vmax_lag_sc:trial_neg_inv_sc:h_HippAntL_neg +
## (1 | id/run)
## Data: df
##
## REML criterion at convergence: 61513.7
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.5985 -0.5438 -0.0677 0.5055 4.6481
##
## Random effects:
## Groups Name Variance Std.Dev.
## run:id (Intercept) 0.05548 0.2355
## id (Intercept) 0.04396 0.2097
## Residual 0.53299 0.7301
## Number of obs: 27253, groups: run:id, 560; id, 70
##
## Fixed effects:
## Estimate Std. Error df
## (Intercept) 1.045e-01 3.230e-02 7.056e+01
## trial_neg_inv_sc 9.945e-03 1.272e-02 2.667e+04
## rt_lag_sc 4.589e-01 9.596e-03 2.721e+04
## rt_vmax_lag_sc 1.087e-01 9.746e-03 2.427e+04
## last_outcomeOmission -1.741e-01 1.210e-02 2.692e+04
## v_max_wi_lag 6.720e-03 6.902e-03 2.680e+04
## v_entropy_wi 2.165e-02 6.932e-03 2.673e+04
## h_HippAntL_neg 7.536e-03 1.394e-01 6.918e+01
## trial_neg_inv_sc:rt_lag_sc -7.907e-02 1.042e-02 2.676e+04
## trial_neg_inv_sc:rt_vmax_lag_sc 7.129e-02 1.093e-02 2.716e+04
## trial_neg_inv_sc:last_outcomeOmission 1.411e-02 1.829e-02 2.668e+04
## trial_neg_inv_sc:v_max_wi_lag 1.454e-02 6.590e-03 2.712e+04
## trial_neg_inv_sc:v_entropy_wi -6.476e-03 7.361e-03 2.718e+04
## trial_neg_inv_sc:h_HippAntL_neg -1.484e-02 4.392e-02 2.691e+04
## rt_lag_sc:rt_vmax_lag_sc 9.450e-04 5.141e-03 2.644e+04
## rt_lag_sc:last_outcomeOmission -3.886e-01 1.280e-02 2.703e+04
## rt_lag_sc:v_max_wi_lag -3.150e-02 6.123e-03 2.694e+04
## rt_lag_sc:v_entropy_wi -1.871e-02 6.114e-03 2.682e+04
## rt_lag_sc:h_HippAntL_neg 8.869e-02 4.081e-02 2.689e+04
## rt_vmax_lag_sc:last_outcomeOmission 7.157e-02 1.181e-02 2.700e+04
## rt_vmax_lag_sc:v_max_wi_lag 7.682e-02 5.928e-03 2.703e+04
## rt_vmax_lag_sc:v_entropy_wi 7.577e-03 6.100e-03 2.701e+04
## rt_vmax_lag_sc:h_HippAntL_neg 1.260e-01 3.809e-02 2.027e+04
## last_outcomeOmission:v_max_wi_lag -1.892e-02 1.011e-02 2.680e+04
## last_outcomeOmission:v_entropy_wi -3.191e-02 1.041e-02 2.677e+04
## last_outcomeOmission:h_HippAntL_neg 1.679e-01 5.137e-02 2.702e+04
## v_max_wi_lag:v_entropy_wi 3.304e-03 4.710e-03 2.718e+04
## v_max_wi_lag:h_HippAntL_neg -4.427e-02 2.393e-02 2.665e+04
## v_entropy_wi:h_HippAntL_neg 2.933e-02 2.462e-02 2.658e+04
## rt_lag_sc:last_outcomeOmission:h_HippAntL_neg 6.079e-02 4.723e-02 2.712e+04
## trial_neg_inv_sc:rt_vmax_lag_sc:h_HippAntL_neg 1.249e-01 4.529e-02 2.717e+04
## t value Pr(>|t|)
## (Intercept) 3.235 0.001851 **
## trial_neg_inv_sc 0.782 0.434250
## rt_lag_sc 47.829 < 2e-16 ***
## rt_vmax_lag_sc 11.157 < 2e-16 ***
## last_outcomeOmission -14.395 < 2e-16 ***
## v_max_wi_lag 0.974 0.330301
## v_entropy_wi 3.124 0.001789 **
## h_HippAntL_neg 0.054 0.957049
## trial_neg_inv_sc:rt_lag_sc -7.587 3.38e-14 ***
## trial_neg_inv_sc:rt_vmax_lag_sc 6.520 7.16e-11 ***
## trial_neg_inv_sc:last_outcomeOmission 0.771 0.440602
## trial_neg_inv_sc:v_max_wi_lag 2.207 0.027341 *
## trial_neg_inv_sc:v_entropy_wi -0.880 0.379024
## trial_neg_inv_sc:h_HippAntL_neg -0.338 0.735380
## rt_lag_sc:rt_vmax_lag_sc 0.184 0.854163
## rt_lag_sc:last_outcomeOmission -30.352 < 2e-16 ***
## rt_lag_sc:v_max_wi_lag -5.146 2.68e-07 ***
## rt_lag_sc:v_entropy_wi -3.059 0.002220 **
## rt_lag_sc:h_HippAntL_neg 2.174 0.029751 *
## rt_vmax_lag_sc:last_outcomeOmission 6.060 1.38e-09 ***
## rt_vmax_lag_sc:v_max_wi_lag 12.960 < 2e-16 ***
## rt_vmax_lag_sc:v_entropy_wi 1.242 0.214182
## rt_vmax_lag_sc:h_HippAntL_neg 3.309 0.000939 ***
## last_outcomeOmission:v_max_wi_lag -1.871 0.061332 .
## last_outcomeOmission:v_entropy_wi -3.066 0.002173 **
## last_outcomeOmission:h_HippAntL_neg 3.268 0.001085 **
## v_max_wi_lag:v_entropy_wi 0.701 0.483088
## v_max_wi_lag:h_HippAntL_neg -1.850 0.064300 .
## v_entropy_wi:h_HippAntL_neg 1.191 0.233562
## rt_lag_sc:last_outcomeOmission:h_HippAntL_neg 1.287 0.198040
## trial_neg_inv_sc:rt_vmax_lag_sc:h_HippAntL_neg 2.758 0.005815 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation matrix not shown by default, as p = 31 > 12.
## Use print(x, correlation=TRUE) or
## vcov(x) if you need it
Anova(model_AH_beta_fMRI, '3')
## Analysis of Deviance Table (Type III Wald chisquare tests)
##
## Response: rt_csv_sc
## Chisq Df Pr(>Chisq)
## (Intercept) 10.4672 1 0.0012151 **
## trial_neg_inv_sc 0.6114 1 0.4342432
## rt_lag_sc 2287.6118 1 < 2.2e-16 ***
## rt_vmax_lag_sc 124.4858 1 < 2.2e-16 ***
## last_outcome 207.2120 1 < 2.2e-16 ***
## v_max_wi_lag 0.9478 1 0.3302922
## v_entropy_wi 9.7564 1 0.0017870 **
## h_HippAntL_neg 0.0029 1 0.9568936
## trial_neg_inv_sc:rt_lag_sc 57.5633 1 3.273e-14 ***
## trial_neg_inv_sc:rt_vmax_lag_sc 42.5095 1 7.034e-11 ***
## trial_neg_inv_sc:last_outcome 0.5947 1 0.4405956
## trial_neg_inv_sc:v_max_wi_lag 4.8697 1 0.0273328 *
## trial_neg_inv_sc:v_entropy_wi 0.7739 1 0.3790166
## trial_neg_inv_sc:h_HippAntL_neg 0.1142 1 0.7353772
## rt_lag_sc:rt_vmax_lag_sc 0.0338 1 0.8541618
## rt_lag_sc:last_outcome 921.2556 1 < 2.2e-16 ***
## rt_lag_sc:v_max_wi_lag 26.4778 1 2.666e-07 ***
## rt_lag_sc:v_entropy_wi 9.3596 1 0.0022182 **
## rt_lag_sc:h_HippAntL_neg 4.7241 1 0.0297424 *
## rt_vmax_lag_sc:last_outcome 36.7272 1 1.359e-09 ***
## rt_vmax_lag_sc:v_max_wi_lag 167.9494 1 < 2.2e-16 ***
## rt_vmax_lag_sc:v_entropy_wi 1.5430 1 0.2141709
## rt_vmax_lag_sc:h_HippAntL_neg 10.9477 1 0.0009372 ***
## last_outcome:v_max_wi_lag 3.5013 1 0.0613212 .
## last_outcome:v_entropy_wi 9.3990 1 0.0021710 **
## last_outcome:h_HippAntL_neg 10.6783 1 0.0010840 **
## v_max_wi_lag:v_entropy_wi 0.4919 1 0.4830818
## v_max_wi_lag:h_HippAntL_neg 3.4231 1 0.0642889 .
## v_entropy_wi:h_HippAntL_neg 1.4191 1 0.2335510
## rt_lag_sc:last_outcome:h_HippAntL_neg 1.6569 1 0.1980285
## trial_neg_inv_sc:rt_vmax_lag_sc:h_HippAntL_neg 7.6079 1 0.0058113 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
# include rt_lag and rt_vmax as random
model_AH_beta_fMRI_random <- lmer(rt_csv_sc ~ (trial_neg_inv_sc + rt_lag_sc + rt_vmax_lag_sc + last_outcome +
v_max_wi_lag + v_entropy_wi + h_HippAntL_neg)^2 +
rt_lag_sc:last_outcome:h_HippAntL_neg +
rt_vmax_lag_sc:trial_neg_inv_sc:h_HippAntL_neg +
(rt_lag_sc + rt_vmax_lag_sc|id/run), df)
## Warning in checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
## Model failed to converge with max|grad| = 0.0046599 (tol = 0.002, component 1)
# in case model no longer converges with more random effects,
# re-run in a loop until it does
while (any(grepl("failed to converge", model_AH_beta_fMRI_random@optinfo$conv$lme4$messages) )) {
print(model_AH_beta_fMRI_random@optinfo$conv$lme4$conv)
ss <- getME(model_AH_beta_fMRI_random,c("theta","fixef"))
model_AH_beta_fMRI_random <- update(model_AH_beta_fMRI_random, start=ss)}
## NULL
summary(model_AH_beta_fMRI_random)
## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: rt_csv_sc ~ (trial_neg_inv_sc + rt_lag_sc + rt_vmax_lag_sc +
## last_outcome + v_max_wi_lag + v_entropy_wi + h_HippAntL_neg)^2 +
## rt_lag_sc:last_outcome:h_HippAntL_neg + rt_vmax_lag_sc:trial_neg_inv_sc:h_HippAntL_neg +
## (rt_lag_sc + rt_vmax_lag_sc | id/run)
## Data: df
##
## REML criterion at convergence: 59926.9
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.6978 -0.5244 -0.0596 0.4711 4.7731
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## run:id (Intercept) 0.031689 0.17801
## rt_lag_sc 0.014588 0.12078 -0.02
## rt_vmax_lag_sc 0.006641 0.08149 0.07 -0.26
## id (Intercept) 0.035720 0.18900
## rt_lag_sc 0.050606 0.22496 -0.03
## rt_vmax_lag_sc 0.003225 0.05679 0.06 -0.05
## Residual 0.493556 0.70254
## Number of obs: 27253, groups: run:id, 560; id, 70
##
## Fixed effects:
## Estimate Std. Error df
## (Intercept) 1.378e-01 2.904e-02 6.836e+01
## trial_neg_inv_sc -2.550e-03 1.254e-02 2.337e+04
## rt_lag_sc 4.991e-01 3.342e-02 7.364e+01
## rt_vmax_lag_sc 7.072e-02 1.408e-02 9.530e+01
## last_outcomeOmission -1.782e-01 1.172e-02 2.453e+04
## v_max_wi_lag 1.736e-02 6.786e-03 2.533e+04
## v_entropy_wi 2.715e-02 6.822e-03 2.574e+04
## h_HippAntL_neg -7.762e-02 1.257e-01 6.781e+01
## trial_neg_inv_sc:rt_lag_sc -6.517e-02 1.030e-02 2.552e+04
## trial_neg_inv_sc:rt_vmax_lag_sc 6.555e-02 1.086e-02 2.020e+04
## trial_neg_inv_sc:last_outcomeOmission 1.929e-02 1.782e-02 2.631e+04
## trial_neg_inv_sc:v_max_wi_lag 8.437e-03 6.446e-03 2.594e+04
## trial_neg_inv_sc:v_entropy_wi -9.060e-03 7.203e-03 2.575e+04
## trial_neg_inv_sc:h_HippAntL_neg 1.763e-02 4.365e-02 2.253e+04
## rt_lag_sc:rt_vmax_lag_sc -4.424e-02 5.982e-03 8.049e+03
## rt_lag_sc:last_outcomeOmission -3.896e-01 1.282e-02 2.198e+04
## rt_lag_sc:v_max_wi_lag -6.111e-03 6.094e-03 2.448e+04
## rt_lag_sc:v_entropy_wi -2.516e-02 6.050e-03 2.381e+04
## rt_lag_sc:h_HippAntL_neg 1.876e-01 1.452e-01 7.399e+01
## rt_vmax_lag_sc:last_outcomeOmission 7.447e-02 1.172e-02 1.858e+04
## rt_vmax_lag_sc:v_max_wi_lag 5.454e-02 5.920e-03 2.272e+04
## rt_vmax_lag_sc:v_entropy_wi 2.041e-02 6.061e-03 2.283e+04
## rt_vmax_lag_sc:h_HippAntL_neg 9.606e-02 5.823e-02 7.962e+01
## last_outcomeOmission:v_max_wi_lag -2.396e-02 9.813e-03 2.665e+04
## last_outcomeOmission:v_entropy_wi -3.684e-02 1.014e-02 2.654e+04
## last_outcomeOmission:h_HippAntL_neg 1.690e-01 4.968e-02 2.525e+04
## v_max_wi_lag:v_entropy_wi 8.170e-03 4.590e-03 2.655e+04
## v_max_wi_lag:h_HippAntL_neg -3.677e-02 2.353e-02 2.396e+04
## v_entropy_wi:h_HippAntL_neg 3.543e-02 2.430e-02 2.459e+04
## rt_lag_sc:last_outcomeOmission:h_HippAntL_neg 4.805e-02 4.646e-02 2.588e+04
## trial_neg_inv_sc:rt_vmax_lag_sc:h_HippAntL_neg 9.446e-02 4.534e-02 2.098e+04
## t value Pr(>|t|)
## (Intercept) 4.745 1.10e-05 ***
## trial_neg_inv_sc -0.203 0.838883
## rt_lag_sc 14.934 < 2e-16 ***
## rt_vmax_lag_sc 5.022 2.38e-06 ***
## last_outcomeOmission -15.209 < 2e-16 ***
## v_max_wi_lag 2.559 0.010504 *
## v_entropy_wi 3.979 6.93e-05 ***
## h_HippAntL_neg -0.617 0.539018
## trial_neg_inv_sc:rt_lag_sc -6.330 2.49e-10 ***
## trial_neg_inv_sc:rt_vmax_lag_sc 6.038 1.59e-09 ***
## trial_neg_inv_sc:last_outcomeOmission 1.083 0.279008
## trial_neg_inv_sc:v_max_wi_lag 1.309 0.190566
## trial_neg_inv_sc:v_entropy_wi -1.258 0.208466
## trial_neg_inv_sc:h_HippAntL_neg 0.404 0.686287
## rt_lag_sc:rt_vmax_lag_sc -7.397 1.54e-13 ***
## rt_lag_sc:last_outcomeOmission -30.374 < 2e-16 ***
## rt_lag_sc:v_max_wi_lag -1.003 0.315933
## rt_lag_sc:v_entropy_wi -4.159 3.21e-05 ***
## rt_lag_sc:h_HippAntL_neg 1.292 0.200475
## rt_vmax_lag_sc:last_outcomeOmission 6.355 2.14e-10 ***
## rt_vmax_lag_sc:v_max_wi_lag 9.213 < 2e-16 ***
## rt_vmax_lag_sc:v_entropy_wi 3.368 0.000759 ***
## rt_vmax_lag_sc:h_HippAntL_neg 1.650 0.102968
## last_outcomeOmission:v_max_wi_lag -2.441 0.014646 *
## last_outcomeOmission:v_entropy_wi -3.634 0.000279 ***
## last_outcomeOmission:h_HippAntL_neg 3.401 0.000673 ***
## v_max_wi_lag:v_entropy_wi 1.780 0.075076 .
## v_max_wi_lag:h_HippAntL_neg -1.562 0.118201
## v_entropy_wi:h_HippAntL_neg 1.458 0.144953
## rt_lag_sc:last_outcomeOmission:h_HippAntL_neg 1.034 0.301006
## trial_neg_inv_sc:rt_vmax_lag_sc:h_HippAntL_neg 2.083 0.037222 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation matrix not shown by default, as p = 31 > 12.
## Use print(x, correlation=TRUE) or
## vcov(x) if you need it
Anova(model_AH_beta_fMRI_random, '3')
## Analysis of Deviance Table (Type III Wald chisquare tests)
##
## Response: rt_csv_sc
## Chisq Df Pr(>Chisq)
## (Intercept) 22.5191 1 2.081e-06 ***
## trial_neg_inv_sc 0.0413 1 0.8388811
## rt_lag_sc 223.0357 1 < 2.2e-16 ***
## rt_vmax_lag_sc 25.2165 1 5.124e-07 ***
## last_outcome 231.3099 1 < 2.2e-16 ***
## v_max_wi_lag 6.5483 1 0.0104984 *
## v_entropy_wi 15.8358 1 6.908e-05 ***
## h_HippAntL_neg 0.3812 1 0.5369496
## trial_neg_inv_sc:rt_lag_sc 40.0717 1 2.448e-10 ***
## trial_neg_inv_sc:rt_vmax_lag_sc 36.4601 1 1.558e-09 ***
## trial_neg_inv_sc:last_outcome 1.1720 1 0.2789983
## trial_neg_inv_sc:v_max_wi_lag 1.7133 1 0.1905544
## trial_neg_inv_sc:v_entropy_wi 1.5821 1 0.2084545
## trial_neg_inv_sc:h_HippAntL_neg 0.1631 1 0.6862836
## rt_lag_sc:rt_vmax_lag_sc 54.7084 1 1.398e-13 ***
## rt_lag_sc:last_outcome 922.5658 1 < 2.2e-16 ***
## rt_lag_sc:v_max_wi_lag 1.0057 1 0.3159233
## rt_lag_sc:v_entropy_wi 17.2980 1 3.195e-05 ***
## rt_lag_sc:h_HippAntL_neg 1.6685 1 0.1964549
## rt_vmax_lag_sc:last_outcome 40.3805 1 2.090e-10 ***
## rt_vmax_lag_sc:v_max_wi_lag 84.8762 1 < 2.2e-16 ***
## rt_vmax_lag_sc:v_entropy_wi 11.3431 1 0.0007573 ***
## rt_vmax_lag_sc:h_HippAntL_neg 2.7211 1 0.0990282 .
## last_outcome:v_max_wi_lag 5.9594 1 0.0146392 *
## last_outcome:v_entropy_wi 13.2092 1 0.0002786 ***
## last_outcome:h_HippAntL_neg 11.5661 1 0.0006717 ***
## v_max_wi_lag:v_entropy_wi 3.1687 1 0.0750643 .
## v_max_wi_lag:h_HippAntL_neg 2.4412 1 0.1181876
## v_entropy_wi:h_HippAntL_neg 2.1247 1 0.1449399
## rt_lag_sc:last_outcome:h_HippAntL_neg 1.0698 1 0.3009959
## trial_neg_inv_sc:rt_vmax_lag_sc:h_HippAntL_neg 4.3408 1 0.0372094 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
########
# out-of-session replication with MEG behavioral data
mmodel_AH_beta_MEG_session <- lmer(rt_csv_sc ~ (trial_neg_inv_sc + rt_lag_sc + rt_vmax_lag_sc + last_outcome +
v_max_wi_lag + v_entropy_wi + h_HippAntL_neg)^2 +
rt_lag_sc:last_outcome:h_HippAntL_neg +
rt_vmax_lag_sc:trial_neg_inv_sc:h_HippAntL_neg +
(1|id/run), mdf)
summary(mmodel_AH_beta_MEG_session)
## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: rt_csv_sc ~ (trial_neg_inv_sc + rt_lag_sc + rt_vmax_lag_sc +
## last_outcome + v_max_wi_lag + v_entropy_wi + h_HippAntL_neg)^2 +
## rt_lag_sc:last_outcome:h_HippAntL_neg + rt_vmax_lag_sc:trial_neg_inv_sc:h_HippAntL_neg +
## (1 | id/run)
## Data: mdf
##
## REML criterion at convergence: 68213.8
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.4755 -0.5003 -0.0409 0.4729 5.0646
##
## Random effects:
## Groups Name Variance Std.Dev.
## run:id (Intercept) 0.05521 0.2350
## id (Intercept) 0.03880 0.1970
## Residual 0.47242 0.6873
## Number of obs: 31999, groups: run:id, 519; id, 65
##
## Fixed effects:
## Estimate Std. Error df
## (Intercept) 9.489e-02 3.155e-02 6.507e+01
## trial_neg_inv_sc 6.394e-02 1.101e-02 3.147e+04
## rt_lag_sc 4.685e-01 9.273e-03 3.196e+04
## rt_vmax_lag_sc 1.478e-01 9.001e-03 3.141e+04
## last_outcomeOmission -1.397e-01 1.015e-02 3.191e+04
## v_max_wi_lag 2.012e-04 6.113e-03 3.162e+04
## v_entropy_wi 3.824e-02 5.950e-03 3.149e+04
## h_HippAntL_neg 4.781e-02 1.390e-01 6.390e+01
## trial_neg_inv_sc:rt_lag_sc -3.631e-03 8.883e-03 3.149e+04
## trial_neg_inv_sc:rt_vmax_lag_sc 4.645e-02 9.343e-03 3.184e+04
## trial_neg_inv_sc:last_outcomeOmission 1.690e-02 1.544e-02 3.151e+04
## trial_neg_inv_sc:v_max_wi_lag -3.476e-03 5.224e-03 3.172e+04
## trial_neg_inv_sc:v_entropy_wi -2.770e-03 6.452e-03 3.180e+04
## trial_neg_inv_sc:h_HippAntL_neg 3.670e-02 3.837e-02 3.162e+04
## rt_lag_sc:rt_vmax_lag_sc -2.066e-02 4.758e-03 3.073e+04
## rt_lag_sc:last_outcomeOmission -3.780e-01 1.190e-02 3.180e+04
## rt_lag_sc:v_max_wi_lag -1.699e-02 5.478e-03 3.171e+04
## rt_lag_sc:v_entropy_wi 3.777e-02 5.678e-03 3.152e+04
## rt_lag_sc:h_HippAntL_neg 3.829e-02 3.653e-02 3.179e+04
## rt_vmax_lag_sc:last_outcomeOmission 5.954e-02 1.074e-02 3.163e+04
## rt_vmax_lag_sc:v_max_wi_lag 7.850e-02 5.216e-03 3.175e+04
## rt_vmax_lag_sc:v_entropy_wi -2.367e-02 5.631e-03 3.160e+04
## rt_vmax_lag_sc:h_HippAntL_neg 6.850e-02 3.218e-02 2.916e+04
## last_outcomeOmission:v_max_wi_lag -5.889e-03 8.787e-03 3.154e+04
## last_outcomeOmission:v_entropy_wi -4.317e-02 9.025e-03 3.151e+04
## last_outcomeOmission:h_HippAntL_neg 5.938e-02 4.493e-02 3.189e+04
## v_max_wi_lag:v_entropy_wi 1.124e-03 4.130e-03 3.178e+04
## v_max_wi_lag:h_HippAntL_neg -1.407e-02 2.092e-02 3.139e+04
## v_entropy_wi:h_HippAntL_neg 1.871e-02 2.126e-02 3.137e+04
## rt_lag_sc:last_outcomeOmission:h_HippAntL_neg 1.984e-01 4.199e-02 3.190e+04
## trial_neg_inv_sc:rt_vmax_lag_sc:h_HippAntL_neg 1.267e-01 3.694e-02 3.188e+04
## t value Pr(>|t|)
## (Intercept) 3.007 0.003745 **
## trial_neg_inv_sc 5.805 6.48e-09 ***
## rt_lag_sc 50.527 < 2e-16 ***
## rt_vmax_lag_sc 16.417 < 2e-16 ***
## last_outcomeOmission -13.761 < 2e-16 ***
## v_max_wi_lag 0.033 0.973745
## v_entropy_wi 6.428 1.31e-10 ***
## h_HippAntL_neg 0.344 0.732033
## trial_neg_inv_sc:rt_lag_sc -0.409 0.682741
## trial_neg_inv_sc:rt_vmax_lag_sc 4.972 6.66e-07 ***
## trial_neg_inv_sc:last_outcomeOmission 1.094 0.273796
## trial_neg_inv_sc:v_max_wi_lag -0.665 0.505804
## trial_neg_inv_sc:v_entropy_wi -0.429 0.667722
## trial_neg_inv_sc:h_HippAntL_neg 0.956 0.338916
## rt_lag_sc:rt_vmax_lag_sc -4.342 1.42e-05 ***
## rt_lag_sc:last_outcomeOmission -31.762 < 2e-16 ***
## rt_lag_sc:v_max_wi_lag -3.102 0.001922 **
## rt_lag_sc:v_entropy_wi 6.651 2.96e-11 ***
## rt_lag_sc:h_HippAntL_neg 1.048 0.294645
## rt_vmax_lag_sc:last_outcomeOmission 5.544 2.98e-08 ***
## rt_vmax_lag_sc:v_max_wi_lag 15.050 < 2e-16 ***
## rt_vmax_lag_sc:v_entropy_wi -4.203 2.64e-05 ***
## rt_vmax_lag_sc:h_HippAntL_neg 2.129 0.033289 *
## last_outcomeOmission:v_max_wi_lag -0.670 0.502777
## last_outcomeOmission:v_entropy_wi -4.783 1.74e-06 ***
## last_outcomeOmission:h_HippAntL_neg 1.322 0.186248
## v_max_wi_lag:v_entropy_wi 0.272 0.785565
## v_max_wi_lag:h_HippAntL_neg -0.673 0.501099
## v_entropy_wi:h_HippAntL_neg 0.880 0.378882
## rt_lag_sc:last_outcomeOmission:h_HippAntL_neg 4.724 2.33e-06 ***
## trial_neg_inv_sc:rt_vmax_lag_sc:h_HippAntL_neg 3.431 0.000603 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation matrix not shown by default, as p = 31 > 12.
## Use print(x, correlation=TRUE) or
## vcov(x) if you need it
Anova(mmodel_AH_beta_MEG_session, '3')
## Analysis of Deviance Table (Type III Wald chisquare tests)
##
## Response: rt_csv_sc
## Chisq Df Pr(>Chisq)
## (Intercept) 9.0433 1 0.0026366 **
## trial_neg_inv_sc 33.7033 1 6.419e-09 ***
## rt_lag_sc 2553.0241 1 < 2.2e-16 ***
## rt_vmax_lag_sc 269.5255 1 < 2.2e-16 ***
## last_outcome 189.3651 1 < 2.2e-16 ***
## v_max_wi_lag 0.0011 1 0.9737449
## v_entropy_wi 41.3208 1 1.292e-10 ***
## h_HippAntL_neg 0.1183 1 0.7309034
## trial_neg_inv_sc:rt_lag_sc 0.1671 1 0.6827387
## trial_neg_inv_sc:rt_vmax_lag_sc 24.7204 1 6.628e-07 ***
## trial_neg_inv_sc:last_outcome 1.1977 1 0.2737879
## trial_neg_inv_sc:v_max_wi_lag 0.4427 1 0.5057993
## trial_neg_inv_sc:v_entropy_wi 0.1843 1 0.6677196
## trial_neg_inv_sc:h_HippAntL_neg 0.9146 1 0.3389085
## rt_lag_sc:rt_vmax_lag_sc 18.8503 1 1.414e-05 ***
## rt_lag_sc:last_outcome 1008.7994 1 < 2.2e-16 ***
## rt_lag_sc:v_max_wi_lag 9.6244 1 0.0019201 **
## rt_lag_sc:v_entropy_wi 44.2322 1 2.916e-11 ***
## rt_lag_sc:h_HippAntL_neg 1.0983 1 0.2946372
## rt_vmax_lag_sc:last_outcome 30.7333 1 2.960e-08 ***
## rt_vmax_lag_sc:v_max_wi_lag 226.5075 1 < 2.2e-16 ***
## rt_vmax_lag_sc:v_entropy_wi 17.6660 1 2.633e-05 ***
## rt_vmax_lag_sc:h_HippAntL_neg 4.5313 1 0.0332807 *
## last_outcome:v_max_wi_lag 0.4491 1 0.5027718
## last_outcome:v_entropy_wi 22.8767 1 1.727e-06 ***
## last_outcome:h_HippAntL_neg 1.7471 1 0.1862387
## v_max_wi_lag:v_entropy_wi 0.0740 1 0.7855634
## v_max_wi_lag:h_HippAntL_neg 0.4526 1 0.5010938
## v_entropy_wi:h_HippAntL_neg 0.7743 1 0.3788757
## rt_lag_sc:last_outcome:h_HippAntL_neg 22.3138 1 2.315e-06 ***
## trial_neg_inv_sc:rt_vmax_lag_sc:h_HippAntL_neg 11.7692 1 0.0006022 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
# include rt_lag and rt_vmax as random
mmodel_AH_beta_MEG_session_random <- lmer(rt_csv_sc ~ (trial_neg_inv_sc + rt_lag_sc + rt_vmax_lag_sc + last_outcome +
v_max_wi_lag + v_entropy_wi + h_HippAntL_neg)^2 +
rt_lag_sc:last_outcome:h_HippAntL_neg +
rt_vmax_lag_sc:trial_neg_inv_sc:h_HippAntL_neg +
(rt_lag_sc + rt_vmax_lag_sc|id/run), mdf)
## Warning in checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
## Model failed to converge with max|grad| = 0.0122109 (tol = 0.002, component 1)
summary(mmodel_AH_beta_MEG_session_random)
## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: rt_csv_sc ~ (trial_neg_inv_sc + rt_lag_sc + rt_vmax_lag_sc +
## last_outcome + v_max_wi_lag + v_entropy_wi + h_HippAntL_neg)^2 +
## rt_lag_sc:last_outcome:h_HippAntL_neg + rt_vmax_lag_sc:trial_neg_inv_sc:h_HippAntL_neg +
## (rt_lag_sc + rt_vmax_lag_sc | id/run)
## Data: mdf
##
## REML criterion at convergence: 65687.5
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.8646 -0.4760 -0.0405 0.4476 5.3976
##
## Random effects:
## Groups Name Variance Std.Dev. Corr
## run:id (Intercept) 0.030911 0.17582
## rt_lag_sc 0.017430 0.13202 0.12
## rt_vmax_lag_sc 0.007242 0.08510 0.17 -0.15
## id (Intercept) 0.034254 0.18508
## rt_lag_sc 0.048808 0.22092 0.08
## rt_vmax_lag_sc 0.005712 0.07558 -0.07 0.10
## Residual 0.427242 0.65364
## Number of obs: 31999, groups: run:id, 519; id, 65
##
## Fixed effects:
## Estimate Std. Error df
## (Intercept) 1.016e-01 2.915e-02 6.352e+01
## trial_neg_inv_sc 5.582e-02 1.071e-02 3.003e+04
## rt_lag_sc 5.235e-01 3.399e-02 6.870e+01
## rt_vmax_lag_sc 8.338e-02 1.522e-02 8.377e+01
## last_outcomeOmission -1.379e-01 9.782e-03 3.066e+04
## v_max_wi_lag 6.878e-03 5.942e-03 3.070e+04
## v_entropy_wi 4.157e-02 5.794e-03 3.122e+04
## h_HippAntL_neg 4.665e-03 1.291e-01 6.364e+01
## trial_neg_inv_sc:rt_lag_sc 8.836e-03 8.710e-03 2.979e+04
## trial_neg_inv_sc:rt_vmax_lag_sc 3.962e-02 9.171e-03 2.718e+04
## trial_neg_inv_sc:last_outcomeOmission 1.227e-02 1.481e-02 3.145e+04
## trial_neg_inv_sc:v_max_wi_lag -2.514e-03 5.059e-03 3.097e+04
## trial_neg_inv_sc:v_entropy_wi -6.068e-03 6.259e-03 3.089e+04
## trial_neg_inv_sc:h_HippAntL_neg 5.734e-02 3.774e-02 2.718e+04
## rt_lag_sc:rt_vmax_lag_sc -5.595e-02 5.368e-03 1.229e+04
## rt_lag_sc:last_outcomeOmission -3.865e-01 1.175e-02 2.818e+04
## rt_lag_sc:v_max_wi_lag 1.221e-02 5.429e-03 2.877e+04
## rt_lag_sc:v_entropy_wi 1.801e-02 5.596e-03 2.888e+04
## rt_lag_sc:h_HippAntL_neg 1.148e-01 1.494e-01 6.675e+01
## rt_vmax_lag_sc:last_outcomeOmission 7.384e-02 1.065e-02 2.306e+04
## rt_vmax_lag_sc:v_max_wi_lag 4.677e-02 5.191e-03 2.772e+04
## rt_vmax_lag_sc:v_entropy_wi -7.857e-03 5.549e-03 2.770e+04
## rt_vmax_lag_sc:h_HippAntL_neg 4.418e-02 6.313e-02 6.412e+01
## last_outcomeOmission:v_max_wi_lag -2.724e-03 8.429e-03 3.135e+04
## last_outcomeOmission:v_entropy_wi -3.616e-02 8.700e-03 3.129e+04
## last_outcomeOmission:h_HippAntL_neg 3.731e-02 4.341e-02 3.039e+04
## v_max_wi_lag:v_entropy_wi 7.356e-04 3.992e-03 3.165e+04
## v_max_wi_lag:h_HippAntL_neg -1.956e-02 2.045e-02 2.970e+04
## v_entropy_wi:h_HippAntL_neg 7.276e-03 2.096e-02 3.041e+04
## rt_lag_sc:last_outcomeOmission:h_HippAntL_neg 1.890e-01 4.077e-02 3.114e+04
## trial_neg_inv_sc:rt_vmax_lag_sc:h_HippAntL_neg 8.690e-02 3.628e-02 2.914e+04
## t value Pr(>|t|)
## (Intercept) 3.487 0.000893 ***
## trial_neg_inv_sc 5.212 1.88e-07 ***
## rt_lag_sc 15.403 < 2e-16 ***
## rt_vmax_lag_sc 5.477 4.44e-07 ***
## last_outcomeOmission -14.093 < 2e-16 ***
## v_max_wi_lag 1.157 0.247105
## v_entropy_wi 7.175 7.39e-13 ***
## h_HippAntL_neg 0.036 0.971280
## trial_neg_inv_sc:rt_lag_sc 1.014 0.310373
## trial_neg_inv_sc:rt_vmax_lag_sc 4.320 1.57e-05 ***
## trial_neg_inv_sc:last_outcomeOmission 0.829 0.407389
## trial_neg_inv_sc:v_max_wi_lag -0.497 0.619214
## trial_neg_inv_sc:v_entropy_wi -0.969 0.332326
## trial_neg_inv_sc:h_HippAntL_neg 1.519 0.128663
## rt_lag_sc:rt_vmax_lag_sc -10.423 < 2e-16 ***
## rt_lag_sc:last_outcomeOmission -32.896 < 2e-16 ***
## rt_lag_sc:v_max_wi_lag 2.249 0.024497 *
## rt_lag_sc:v_entropy_wi 3.218 0.001291 **
## rt_lag_sc:h_HippAntL_neg 0.769 0.444813
## rt_vmax_lag_sc:last_outcomeOmission 6.930 4.31e-12 ***
## rt_vmax_lag_sc:v_max_wi_lag 9.009 < 2e-16 ***
## rt_vmax_lag_sc:v_entropy_wi -1.416 0.156858
## rt_vmax_lag_sc:h_HippAntL_neg 0.700 0.486602
## last_outcomeOmission:v_max_wi_lag -0.323 0.746573
## last_outcomeOmission:v_entropy_wi -4.157 3.24e-05 ***
## last_outcomeOmission:h_HippAntL_neg 0.860 0.390010
## v_max_wi_lag:v_entropy_wi 0.184 0.853780
## v_max_wi_lag:h_HippAntL_neg -0.956 0.338884
## v_entropy_wi:h_HippAntL_neg 0.347 0.728442
## rt_lag_sc:last_outcomeOmission:h_HippAntL_neg 4.635 3.58e-06 ***
## trial_neg_inv_sc:rt_vmax_lag_sc:h_HippAntL_neg 2.395 0.016613 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation matrix not shown by default, as p = 31 > 12.
## Use print(x, correlation=TRUE) or
## vcov(x) if you need it
## optimizer (nloptwrap) convergence code: 0 (OK)
## Model failed to converge with max|grad| = 0.0122109 (tol = 0.002, component 1)
Anova(mmodel_AH_beta_MEG_session_random, '3')
## Analysis of Deviance Table (Type III Wald chisquare tests)
##
## Response: rt_csv_sc
## Chisq Df Pr(>Chisq)
## (Intercept) 12.1564 1 0.0004892 ***
## trial_neg_inv_sc 27.1689 1 1.864e-07 ***
## rt_lag_sc 237.2618 1 < 2.2e-16 ***
## rt_vmax_lag_sc 29.9997 1 4.321e-08 ***
## last_outcome 198.6073 1 < 2.2e-16 ***
## v_max_wi_lag 1.3396 1 0.2470958
## v_entropy_wi 51.4806 1 7.231e-13 ***
## h_HippAntL_neg 0.0013 1 0.9711671
## trial_neg_inv_sc:rt_lag_sc 1.0291 1 0.3103644
## trial_neg_inv_sc:rt_vmax_lag_sc 18.6629 1 1.560e-05 ***
## trial_neg_inv_sc:last_outcome 0.6864 1 0.4073823
## trial_neg_inv_sc:v_max_wi_lag 0.2470 1 0.6192105
## trial_neg_inv_sc:v_entropy_wi 0.9398 1 0.3323188
## trial_neg_inv_sc:h_HippAntL_neg 2.3087 1 0.1286515
## rt_lag_sc:rt_vmax_lag_sc 108.6456 1 < 2.2e-16 ***
## rt_lag_sc:last_outcome 1082.1785 1 < 2.2e-16 ***
## rt_lag_sc:v_max_wi_lag 5.0596 1 0.0244898 *
## rt_lag_sc:v_entropy_wi 10.3578 1 0.0012893 **
## rt_lag_sc:h_HippAntL_neg 0.5908 1 0.4421021
## rt_vmax_lag_sc:last_outcome 48.0285 1 4.201e-12 ***
## rt_vmax_lag_sc:v_max_wi_lag 81.1708 1 < 2.2e-16 ***
## rt_vmax_lag_sc:v_entropy_wi 2.0044 1 0.1568466
## rt_vmax_lag_sc:h_HippAntL_neg 0.4897 1 0.4840706
## last_outcome:v_max_wi_lag 0.1044 1 0.7465710
## last_outcome:v_entropy_wi 17.2779 1 3.229e-05 ***
## last_outcome:h_HippAntL_neg 0.7389 1 0.3900030
## v_max_wi_lag:v_entropy_wi 0.0340 1 0.8537791
## v_max_wi_lag:h_HippAntL_neg 0.9147 1 0.3388758
## v_entropy_wi:h_HippAntL_neg 0.1205 1 0.7284396
## rt_lag_sc:last_outcome:h_HippAntL_neg 21.4869 1 3.563e-06 ***
## trial_neg_inv_sc:rt_vmax_lag_sc:h_HippAntL_neg 5.7375 1 0.0166063 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
############################
## Emtrends plot for betas -> behavior figure
emodel_condition <- as_tibble(emtrends(model_AH_beta_fMRI, var = "rt_vmax_lag_sc", specs = c("h_HippAntL_neg", "trial_neg_inv_sc"), at = list(h_HippAntL_neg = c(-.1, .37), trial_neg_inv_sc = c(-.7, 0.44)), options = list()))
## Note: D.f. calculations have been disabled because the number of observations exceeds 3000.
## To enable adjustments, add the argument 'pbkrtest.limit = 27253' (or larger)
## [or, globally, 'set emm_options(pbkrtest.limit = 27253)' or larger];
## but be warned that this may result in large computation time and memory use.
## Note: D.f. calculations have been disabled because the number of observations exceeds 3000.
## To enable adjustments, add the argument 'lmerTest.limit = 27253' (or larger)
## [or, globally, 'set emm_options(lmerTest.limit = 27253)' or larger];
## but be warned that this may result in large computation time and memory use.
emodel_condition$study = 'fMRI'
emodel_RT_Vmax <- as_tibble(emtrends(mmodel_AH_beta_MEG_session, var = "rt_vmax_lag_sc", specs = c("h_HippAntL_neg", "trial_neg_inv_sc"), at = list(h_HippAntL_neg = c(-.1, .37), trial_neg_inv_sc = c(-.7, 0.44)), options = list()))
## Note: D.f. calculations have been disabled because the number of observations exceeds 3000.
## To enable adjustments, add the argument 'pbkrtest.limit = 31999' (or larger)
## [or, globally, 'set emm_options(pbkrtest.limit = 31999)' or larger];
## but be warned that this may result in large computation time and memory use.
## Note: D.f. calculations have been disabled because the number of observations exceeds 3000.
## To enable adjustments, add the argument 'lmerTest.limit = 31999' (or larger)
## [or, globally, 'set emm_options(lmerTest.limit = 31999)' or larger];
## but be warned that this may result in large computation time and memory use.
emodel_RT_Vmax$study = 'Replication'
em4 <- rbind(emodel_condition, emodel_RT_Vmax)
ggplot(em4, aes(x=as.factor(trial_neg_inv_sc), y=rt_vmax_lag_sc.trend, ymin=asymp.LCL, ymax=asymp.UCL, color=as.factor(h_HippAntL_neg))) +
geom_point(position = position_dodge(width = .6), size=2.5) +
geom_errorbar(position = position_dodge(width=0.6), width=0.4, size=0.9) +
theme_bw(base_size=12) + facet_wrap(~study)+ ylab("Convergence on\nbest RT (AU)") +
scale_color_manual("AH global max\nresponse", values=c("#403202", "#e2b407"), labels = c("10th %ile", "90th %ile")) +
labs(shape = "AH global max\nresponse") +
theme(axis.title.x=element_blank(), panel.grid.major.x=element_blank(),
axis.text=element_text(size=8.5, color="grey10")) +
scale_x_discrete(name ="Trial", labels=c("-0.7" = "5", "0.44" = "50")) + scale_y_continuous(limits = c(.05, .3))
# save(file = 'vhd_u_meg_models.Rdata', list = ls(all.names = TRUE))
# load(file = 'vhd_u_meg_models.Rdata')