Hyperreactivity to uncertainty is a key feature of subjective cognitive impairment

Generalised mixed effects models investigating the effect of the group (SCI vs. controls) on performance. Models were specified as follows: Response variable ~ 1 + group*η + group*R₀ + η_s*R₀ + group:η_s :R₀ + (1 + η_s*R₀ |participant). SCI: subjective cognitive impairment. R₀ : initial credit reserve. η_s : sampling cost. EE: uncertainty before committing to decisions.

Number of samples obtained per condition. R₀ : initial credit reserve. η_s : sampling cost.

Deviation from optimal (s−s*) for each condition. R₀ : initial credit reserve. η_s : sampling cost. : number of samples obtained. s*: optimal number of samples.

Generalised mixed effects models investigating the effects ofthe group (SCI vs. controls) and affective burden on uncertainty estimation. Models were specified as follows: Subjective Uncertainty ~ 1 + group* EE+ (1 + EE |participant) + (1 |trial). Subjective Uncertainty ~ 1 + Burden* EE+ age + ACE-III +(1 + EE |participant) + (1 |trial). EE: Experimentally defined expected error. ACE-III: Addenbrook’s Cognitive Examination score.

Generalised mixed effects models investigating effects of group (SCI vs. controls) and affective burden effect on passive decision making under uncertainty. Models were specified as follows: choice ~ 1 + group*R + group*EE + R*EE + group:R:EE + (1 + R*EE |participant). Choice ~ 1 + Burden*R + Burden*EE + R*EE + Burden:R:EE + Age + ACE-III (1 + R*EE|participant). R: reward on offer. EE: expected localisation error. SCI: subjective cognitive impairment group. ACE-III: Addenbrook’s Cognitive Examination score.

Generalised mixed effects models investigating the effect of the group (SCI vs. controls) on sampling speed (ISI) and efficiency (α). Models were specified as follows: Inter-sampling Interval (ISI) ~ 1 + group*η_s + group*R₀ + η_s*R₀ + group:η_s :R₀ + (1 + η_s*R₀ |participant); Information extraction rate (α) ~ 1 + group*η_s + group*R₀ + η_s*R₀ + group:η_s :R₀ + (1 + η_s*R₀ |participant).

R₀ : initial credit reserve. η_s : sampling cost.

Speed efficiency trade-off. Models were specified as follows: α ~ 1 + ISI + (1 +ISI |participant) + (1 + ISI |condition) + (1 |trial).

PaHC: para-hippocampus. Hipp: hippocampus. IC: insular cortex. unc: uncorrected. FWE: family-wise error. FDR: false discovery rate.

PaHC: parahippocampus. Hipp: hippocampus. IC: insular cortex. unc: uncorrected. FWE: family-wise error. FDR: false discovery rate.

(a) Six motion parameters were used during realignment procedure for rfMRI processing. These correspond to six timeseries containing three transnational and three rotational parameters over time for each subject. None of these parameters was significantly different between SCI and controls groups (all ${\displaystyle p_{corr}>0.61}$). (b) Five quality control estimates were used during preprocessing of neuroimaging data (Whitfield-Gabrieli and Nieto-Castanon, 2012). These included number of valid scans after scrubbing procedure, mean and maximum motion (extracted from the six parameters above), mean and maximum global signal change. None of these parameters was significantly different between the two groups (all ${\displaystyle p_{corr}>0.16}$ ). Based on mean motion and mean global signal changes, four potential outliers (three SCI participants and one control with values above or below ${\displaystyle Q3+1.5IQR}$) were identified. A second version of neuroimaging analysis was performed with these participants excluded (Supplementary file 10). There were no changes to the results or conclusions made in the paper. These findings suggest that rfMRI differences between SCI participants and controls are unlikely due to motion artifacts. Mean and max motion was calculated based on Power et al., 2012. Error bars show ± 95% CI.

Across study participants, no correlation was found between mean motion (or global signal change) and hyperreactivity to uncertainty (${\displaystyle ISI}$ or Deviation from optimal) or affective burden (all ${\displaystyle p_{corr}=1}$ ). Specifically, no correlation between ${\displaystyle ISI}$ (the measure that correlates with insular-hippocampal connectivity) and these quality control measures (mean motion and mean GS change) across SCI participants (${\displaystyle p=0.13\mathrm{\&}p=0.49}$ , respectively). These findings suggest that correlation between ${\displaystyle ISI}$ and insular-hippocampal connectivity is unlikely due to motion artifacts. Correlations were controlled for age and gender. Error bars show ± 95% CI.