Anna_optionstudy

RT LMMs

Is RT predicted by difficulty, options or task (correct for option chosen)?

–> YES - slower due to difficulty, options, task, also effects of bestopt and diff:options:task

lmer.RT_diffoptstask_corr <- lmer(scale(RT) ~ scale(Difficulty)*no.options*Task + scale(bestopt_std)  + (1 | ID), data = dymo.alldata, control=lmerControl(optimizer='bobyqa', optCtrl=list(maxfun=2e5)))

tab_model(lmer.RT_diffoptstask_corr) #

	scale(RT)
Predictors	Estimates	CI	p
(Intercept)	-0.01	-0.12 – 0.09	0.779
Difficulty	0.03	0.00 – 0.06	0.029
no options [3]	0.43	0.39 – 0.47	<0.001
Task [Reward]	-0.47	-0.51 – -0.43	<0.001
bestopt std	0.39	0.37 – 0.40	<0.001
Difficulty × no options [3]	0.02	-0.01 – 0.06	0.229
Difficulty × Task [Reward]	0.04	-0.00 – 0.08	0.062
no options [3] × Task [Reward]	0.05	-0.00 – 0.11	0.054
(Difficulty × no options [3]) × Task [Reward]	0.07	0.01 – 0.12	0.016
Random Effects
σ²	0.62
τ₀₀ _ID	0.10
ICC	0.14
N _ID	42
Observations	12443
Marginal R² / Conditional R²	0.284 / 0.388

emmip(lmer.RT_diffoptstask_corr, no.options~Difficulty | Task, type = 'response', cov.reduce = FALSE)

Note: D.f. calculations have been disabled because the number of observations exceeds 3000.
To enable adjustments, add the argument 'pbkrtest.limit = 12443' (or larger)
[or, globally, 'set emm_options(pbkrtest.limit = 12443)' 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 = 12443' (or larger)
[or, globally, 'set emm_options(lmerTest.limit = 12443)' or larger];
but be warned that this may result in large computation time and memory use.

Is RT predicted by das total score,difficulty, options or task (correct for bestoption)

–> greater effect of apathy on RT on effort task

–> greater effect of options in more apathetic ppl in reward task (although slower in 2 option trials in effort compared to reward task)

lmer.RT_diffoptsDAStottask_corr <- lmer(scale(RT) ~ Task*scale(Difficulty)*no.options*scale(scoredas_1) + scale(bestopt_std) + (1 | ID), data = dymo.alldata, control=lmerControl(optimizer='bobyqa', optCtrl=list(maxfun=2e5)))

tab_model(lmer.RT_diffoptsDAStottask_corr) #sig effect of task:das, task:options:das

	scale(RT)
Predictors	Estimates	CI	p
(Intercept)	-0.02	-0.12 – 0.09	0.764
Task [Reward]	-0.48	-0.52 – -0.44	<0.001
Difficulty	0.03	0.00 – 0.06	0.028
no options [3]	0.43	0.39 – 0.47	<0.001
scoredas 1	0.07	-0.02 – 0.17	0.123
bestopt std	0.39	0.37 – 0.40	<0.001
Task [Reward] × Difficulty	0.04	-0.00 – 0.08	0.062
Task [Reward] × no options [3]	0.06	-0.00 – 0.11	0.051
Difficulty × no options [3]	0.02	-0.01 – 0.06	0.231
Task [Reward] × scoredas 1	-0.08	-0.12 – -0.04	<0.001
Difficulty × scoredas 1	-0.02	-0.05 – 0.00	0.109
no options [3] × scoredas 1	-0.01	-0.04 – 0.03	0.670
(Task [Reward] × Difficulty) × no options [3]	0.07	0.01 – 0.12	0.016
(Task [Reward] × Difficulty) × scoredas 1	0.04	-0.00 – 0.08	0.057
(Task [Reward] × no options [3]) × scoredas 1	0.06	0.00 – 0.11	0.046
(Difficulty × no options [3]) × scoredas 1	0.01	-0.03 – 0.04	0.605
(Task [Reward] × Difficulty × no options [3]) × scoredas 1	-0.04	-0.09 – 0.02	0.216
Random Effects
σ²	0.62
τ₀₀ _ID	0.10
ICC	0.14
N _ID	42
Observations	12443
Marginal R² / Conditional R²	0.287 / 0.389

emmip(lmer.RT_diffoptsDAStottask_corr, Task~scoredas_1, type = 'response', cov.reduce = FALSE)

Note: D.f. calculations have been disabled because the number of observations exceeds 3000.
To enable adjustments, add the argument 'pbkrtest.limit = 12443' (or larger)
[or, globally, 'set emm_options(pbkrtest.limit = 12443)' 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 = 12443' (or larger)
[or, globally, 'set emm_options(lmerTest.limit = 12443)' or larger];
but be warned that this may result in large computation time and memory use.

NOTE: Results may be misleading due to involvement in interactions

emmip(lmer.RT_diffoptsDAStottask_corr, no.options~scoredas_1 | Task, type = 'response', cov.reduce = FALSE)

Note: D.f. calculations have been disabled because the number of observations exceeds 3000.
To enable adjustments, add the argument 'pbkrtest.limit = 12443' (or larger)
[or, globally, 'set emm_options(pbkrtest.limit = 12443)' 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 = 12443' (or larger)
[or, globally, 'set emm_options(lmerTest.limit = 12443)' or larger];
but be warned that this may result in large computation time and memory use.

NOTE: Results may be misleading due to involvement in interactions

Is RT predicted by das exec score,difficulty, options or task (corrected for bestoption)

–> slower RT with higher DAS

–> particularly slowed in low difficulty tasks with higher das_exec scores

–> particularly slowed in effort compared with reward task

–> the difficulty effects (more slowed in easy compared to hard tasks) mainly seen in effort task

lmer.RT_diffoptsDASextask_corr <- lmer(scale(RT) ~ scale(Difficulty)*no.options*scale(das_executive_1)*Task + scale(bestopt_std) + (1 | ID), data = dymo.alldata, control=lmerControl(optimizer='bobyqa', optCtrl=list(maxfun=2e5)))

tab_model(lmer.RT_diffoptsDASextask_corr) # sig effect of das_exec, diff:das_exec, das_exec:task, and diff:dasexec:task

	scale(RT)
Predictors	Estimates	CI	p
(Intercept)	-0.02	-0.11 – 0.08	0.750
Difficulty	0.03	0.00 – 0.06	0.027
no options [3]	0.43	0.39 – 0.47	<0.001
das executive 1	0.13	0.03 – 0.23	0.011
Task [Reward]	-0.48	-0.52 – -0.44	<0.001
bestopt std	0.39	0.37 – 0.40	<0.001
Difficulty × no options [3]	0.02	-0.01 – 0.06	0.224
Difficulty × das executive 1	-0.05	-0.08 – -0.02	<0.001
no options [3] × das executive 1	0.01	-0.03 – 0.05	0.630
Difficulty × Task [Reward]	0.04	-0.00 – 0.08	0.065
no options [3] × Task [Reward]	0.05	-0.00 – 0.11	0.053
das executive 1 × Task [Reward]	-0.08	-0.12 – -0.04	<0.001
(Difficulty × no options [3]) × das executive 1	-0.01	-0.05 – 0.03	0.550
(Difficulty × no options [3]) × Task [Reward]	0.07	0.01 – 0.12	0.016
(Difficulty × das executive 1) × Task [Reward]	0.05	0.01 – 0.09	0.019
(no options [3] × das executive 1) × Task [Reward]	0.01	-0.05 – 0.06	0.796
(Difficulty × no options [3] × das executive 1) × Task [Reward]	0.03	-0.02 – 0.09	0.279
Random Effects
σ²	0.62
τ₀₀ _ID	0.10
ICC	0.14
N _ID	42
Observations	12443
Marginal R² / Conditional R²	0.296 / 0.392

emmip(lmer.RT_diffoptsDASextask_corr, Difficulty~das_executive_1, type = 'response', cov.reduce = FALSE)

Note: D.f. calculations have been disabled because the number of observations exceeds 3000.
To enable adjustments, add the argument 'pbkrtest.limit = 12443' (or larger)
[or, globally, 'set emm_options(pbkrtest.limit = 12443)' 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 = 12443' (or larger)
[or, globally, 'set emm_options(lmerTest.limit = 12443)' or larger];
but be warned that this may result in large computation time and memory use.

NOTE: Results may be misleading due to involvement in interactions

emmip(lmer.RT_diffoptsDASextask_corr, Task~das_executive_1, type = 'response', cov.reduce = FALSE)

Note: D.f. calculations have been disabled because the number of observations exceeds 3000.
To enable adjustments, add the argument 'pbkrtest.limit = 12443' (or larger)
[or, globally, 'set emm_options(pbkrtest.limit = 12443)' 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 = 12443' (or larger)
[or, globally, 'set emm_options(lmerTest.limit = 12443)' or larger];
but be warned that this may result in large computation time and memory use.

NOTE: Results may be misleading due to involvement in interactions

emmip(lmer.RT_diffoptsDASextask_corr, Difficulty~das_executive_1 | Task, type = 'response', cov.reduce = FALSE)

Note: D.f. calculations have been disabled because the number of observations exceeds 3000.
To enable adjustments, add the argument 'pbkrtest.limit = 12443' (or larger)
[or, globally, 'set emm_options(pbkrtest.limit = 12443)' 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 = 12443' (or larger)
[or, globally, 'set emm_options(lmerTest.limit = 12443)' or larger];
but be warned that this may result in large computation time and memory use.

NOTE: Results may be misleading due to involvement in interactions

Is RT predicted by das auto-activation score,difficulty, options or task (corrected for bestoption)?

–> not really, 4 way interaction present but uninterpretable

lmer.RT_diffoptsDASbehtask_corr <- lmer(scale(RT) ~ scale(Difficulty)*no.options*scale(das_activation_1)*Task + scale(bestopt_std) + (1 | ID), data = dymo.alldata, control=lmerControl(optimizer='bobyqa', optCtrl=list(maxfun=2e5)))

tab_model(lmer.RT_diffoptsDASbehtask_corr) #no effect of das behav, sig effect of das behav:task

	scale(RT)
Predictors	Estimates	CI	p
(Intercept)	-0.01	-0.12 – 0.09	0.784
Difficulty	0.03	0.00 – 0.06	0.031
no options [3]	0.43	0.40 – 0.47	<0.001
das activation 1	0.01	-0.09 – 0.11	0.901
Task [Reward]	-0.47	-0.51 – -0.43	<0.001
bestopt std	0.39	0.37 – 0.40	<0.001
Difficulty × no options [3]	0.02	-0.01 – 0.06	0.232
Difficulty × das activation 1	0.02	-0.01 – 0.04	0.184
no options [3] × das activation 1	-0.03	-0.06 – 0.01	0.132
Difficulty × Task [Reward]	0.04	-0.00 – 0.08	0.054
no options [3] × Task [Reward]	0.05	-0.00 – 0.11	0.054
das activation 1 × Task [Reward]	0.02	-0.02 – 0.07	0.253
(Difficulty × no options [3]) × das activation 1	0.01	-0.03 – 0.05	0.575
(Difficulty × no options [3]) × Task [Reward]	0.07	0.01 – 0.12	0.017
(Difficulty × das activation 1) × Task [Reward]	0.03	-0.01 – 0.06	0.214
(no options [3] × das activation 1) × Task [Reward]	0.06	0.00 – 0.12	0.036
(Difficulty × no options [3] × das activation 1) × Task [Reward]	-0.06	-0.12 – -0.01	0.026
Random Effects
σ²	0.62
τ₀₀ _ID	0.11
ICC	0.15
N _ID	42
Observations	12443
Marginal R² / Conditional R²	0.284 / 0.392

emmip(lmer.RT_diffoptsDASbehtask_corr, Task~das_activation_1 | no.options | Task, type = 'response', cov.reduce = FALSE)

Note: D.f. calculations have been disabled because the number of observations exceeds 3000.
To enable adjustments, add the argument 'pbkrtest.limit = 12443' (or larger)
[or, globally, 'set emm_options(pbkrtest.limit = 12443)' 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 = 12443' (or larger)
[or, globally, 'set emm_options(lmerTest.limit = 12443)' or larger];
but be warned that this may result in large computation time and memory use.

NOTE: Results may be misleading due to involvement in interactions

Accuracy GLMMs

Is accuracy predicted by difficulty, options or task (correct for bestoption)?

–> Significant effects of difficulty, options, task ,(best option)

glmer.Acc_DiffOptsTask_corr <- glmer(accuracy ~ scale(Difficulty)*no.options*Task + scale(bestopt_std) + (1 | ID), family = 'binomial', data = dymo.alldata, control=glmerControl(optimizer='bobyqa', optCtrl=list(maxfun=2e5)))

tab_model(glmer.Acc_DiffOptsTask_corr)

	accuracy
Predictors	Odds Ratios	CI	p
(Intercept)	29.38	21.15 – 40.80	<0.001
Difficulty	0.33	0.27 – 0.40	<0.001
no options [3]	0.70	0.54 – 0.92	0.010
Task [Reward]	2.05	1.40 – 2.98	<0.001
bestopt std	0.73	0.68 – 0.79	<0.001
Difficulty × no options [3]	1.09	0.85 – 1.40	0.489
Difficulty × Task [Reward]	0.97	0.70 – 1.36	0.878
no options [3] × Task [Reward]	1.04	0.64 – 1.69	0.878
(Difficulty × no options [3]) × Task [Reward]	1.17	0.75 – 1.82	0.491
Random Effects
σ²	3.29
τ₀₀ _ID	0.67
ICC	0.17
N _ID	42
Observations	12443
Marginal R² / Conditional R²	0.296 / 0.416

emmip(glmer.Acc_DiffOptsTask_corr, Difficulty~Task, type = 'response', cov.reduce = FALSE)

NOTE: Results may be misleading due to involvement in interactions

Is accuracy predicted by DAS total score, difficulty, options or task (corrected for bestoption)?

–> no significant effects of DAS or interactions

glmer.Acc_DiffOptsDAStotTask_corr <- glmer(accuracy ~ scale(Difficulty)*no.options*scale(scoredas_1)*Task + scale(bestopt_std) + (1 | ID), family = 'binomial', data = dymo.alldata, control=glmerControl(optimizer='bobyqa', optCtrl=list(maxfun=2e5)))

tab_model(glmer.Acc_DiffOptsDAStotTask_corr)

	accuracy
Predictors	Odds Ratios	CI	p
(Intercept)	29.86	21.49 – 41.50	<0.001
Difficulty	0.33	0.27 – 0.40	<0.001
no options [3]	0.69	0.52 – 0.91	0.009
scoredas 1	0.82	0.61 – 1.10	0.178
Task [Reward]	2.14	1.43 – 3.19	<0.001
bestopt std	0.73	0.68 – 0.79	<0.001
Difficulty × no options [3]	1.10	0.85 – 1.42	0.452
Difficulty × scoredas 1	1.06	0.89 – 1.25	0.519
no options [3] × scoredas 1	1.07	0.84 – 1.36	0.569
Difficulty × Task [Reward]	0.94	0.66 – 1.34	0.722
no options [3] × Task [Reward]	1.05	0.63 – 1.76	0.852
scoredas 1 × Task [Reward]	0.82	0.53 – 1.27	0.382
(Difficulty × no options [3]) × scoredas 1	0.98	0.79 – 1.23	0.885
(Difficulty × no options [3]) × Task [Reward]	1.17	0.73 – 1.86	0.511
(Difficulty × scoredas 1) × Task [Reward]	1.25	0.85 – 1.82	0.252
(no options [3] × scoredas 1) × Task [Reward]	0.91	0.53 – 1.59	0.752
(Difficulty × no options [3] × scoredas 1) × Task [Reward]	0.91	0.56 – 1.50	0.723
Random Effects
σ²	3.29
τ₀₀ _ID	0.65
ICC	0.16
N _ID	42
Observations	12443
Marginal R² / Conditional R²	0.321 / 0.433

Is accuracy predicted by DAS exec score, difficulty, options or task (corrected for bestoption)?

-> higher executive apathy assoc with lower accuracy, particularly in difficult tasks

glmer.Acc_DiffOptsDASExTask_corr <- glmer(accuracy ~ scale(Difficulty)*no.options*scale(das_executive_1)*Task + scale(bestopt_std)  + (1 | ID), family = 'binomial', data = dymo.alldata, control=glmerControl(optimizer='bobyqa', optCtrl=list(maxfun=2e5)))

tab_model(glmer.Acc_DiffOptsDASExTask_corr) #sig effects of das_exec and das_exec:difficulty

	accuracy
Predictors	Odds Ratios	CI	p
(Intercept)	34.29	24.11 – 48.76	<0.001
Difficulty	0.29	0.23 – 0.36	<0.001
no options [3]	0.63	0.46 – 0.87	0.005
das executive 1	0.55	0.39 – 0.79	0.001
Task [Reward]	1.87	1.21 – 2.90	0.005
bestopt std	0.73	0.68 – 0.79	<0.001
Difficulty × no options [3]	1.19	0.89 – 1.59	0.231
Difficulty × das executive 1	1.34	1.07 – 1.68	0.010
no options [3] × das executive 1	1.20	0.87 – 1.66	0.258
Difficulty × Task [Reward]	1.06	0.72 – 1.56	0.761
no options [3] × Task [Reward]	1.05	0.61 – 1.81	0.851
das executive 1 × Task [Reward]	1.15	0.74 – 1.79	0.523
(Difficulty × no options [3]) × das executive 1	0.85	0.64 – 1.13	0.254
(Difficulty × no options [3]) × Task [Reward]	1.12	0.69 – 1.82	0.645
(Difficulty × das executive 1) × Task [Reward]	0.83	0.56 – 1.21	0.331
(no options [3] × das executive 1) × Task [Reward]	1.43	0.84 – 2.43	0.189
(Difficulty × no options [3] × das executive 1) × Task [Reward]	0.90	0.56 – 1.44	0.648
Random Effects
σ²	3.29
τ₀₀ _ID	0.59
ICC	0.15
N _ID	42
Observations	12443
Marginal R² / Conditional R²	0.343 / 0.443

emmip(glmer.Acc_DiffOptsDASExTask_corr, Difficulty~das_executive_1, type = 'response', cov.reduce = FALSE)

NOTE: Results may be misleading due to involvement in interactions

Is accuracy predicted by DAS behav score, difficulty, options or task (correct for bestoption)?

–>no sig effects of das_behav, there is a three way interaction of options:DAS_behav:task

–> overall difficult to interpret - seemingly greater detrimental effect of options on accuracy in those with behavioural apathy only in reward task

glmer.Acc_DiffOptsDASBehTask_corr <- glmer(accuracy ~ scale(Difficulty)*no.options*scale(das_activation_1)*Task + scale(bestopt_std)  + (1 | ID), family = 'binomial', data = dymo.alldata, control=glmerControl(optimizer='bobyqa', optCtrl=list(maxfun=2e5)))

tab_model(glmer.Acc_DiffOptsDASBehTask_corr)

	accuracy
Predictors	Odds Ratios	CI	p
(Intercept)	29.44	21.20 – 40.90	<0.001
Difficulty	0.33	0.27 – 0.40	<0.001
no options [3]	0.71	0.54 – 0.93	0.012
das activation 1	1.06	0.79 – 1.43	0.681
Task [Reward]	2.06	1.41 – 3.01	<0.001
bestopt std	0.73	0.68 – 0.78	<0.001
Difficulty × no options [3]	1.09	0.85 – 1.40	0.509
Difficulty × das activation 1	0.97	0.82 – 1.13	0.675
no options [3] × das activation 1	1.09	0.87 – 1.38	0.437
Difficulty × Task [Reward]	1.00	0.71 – 1.40	0.993
no options [3] × Task [Reward]	1.07	0.65 – 1.76	0.792
das activation 1 × Task [Reward]	1.04	0.72 – 1.49	0.844
(Difficulty × no options [3]) × das activation 1	1.00	0.81 – 1.24	0.993
(Difficulty × no options [3]) × Task [Reward]	1.09	0.69 – 1.71	0.707
(Difficulty × das activation 1) × Task [Reward]	1.17	0.85 – 1.62	0.330
(no options [3] × das activation 1) × Task [Reward]	0.56	0.35 – 0.89	0.015
(Difficulty × no options [3] × das activation 1) × Task [Reward]	1.13	0.74 – 1.73	0.578
Random Effects
σ²	3.29
τ₀₀ _ID	0.67
ICC	0.17
N _ID	42
Observations	12443
Marginal R² / Conditional R²	0.306 / 0.424

emmip(glmer.Acc_DiffOptsDASBehTask_corr, no.options~das_activation_1 | Task, type = 'response', cov.reduce = FALSE)

NOTE: Results may be misleading due to involvement in interactions