Participant demographic characteristics are reported in .
Characteristics of patients with bipolar disorder and non-clinical controls: means and SDs.
The patients with bipolar disorder were divided into two groups, a ‘stable’ group (6 males, 5 females) and an ‘unstable’ one (7 males, 5 females). Mood score charts from the twenty-three patients were categorized independently by two psychiatrists (blind to imagery scores) on the basis of the visual inspection of the clinical charts. Concordance of scores was 100%. The demographic details of the two groups are shown in (top panel). Compared to those in the stable group, participants in the unstable group had higher scores of depression (BDI), t(21) = 4.31, p < .001, Mdiff = 13.29, 95% CI [6.88, 19.70] and anxiety (STAI), t(21) = 4.08, p = .001, Mdiff = 17.63, 95% CI [8.64, 26.62], and any difference for mania scores did not reach significance t(21) = 1.09, p = .29 (see , top panel).
Characteristics of patients within the stable versus unstable mood group (top panel) and details of their weekly SMS mood ratings data (bottom panel).
In addition to clinical inspection of the charts by the psychiatrists, weekly mood scores were subjected to statistical analysis. The descriptive data for the weekly mood ratings are shown in (bottom panel). Given the non-normal distribution of this data we used a non-parametric Mann Whitney U test (which is based on ranks of all the individual mood scores) to compare stable and unstable groups. There were significant differences between the patient groups based on both the QIDS-SR (Mann Whitney U test: W = 51,022, p < .001) and ASRM (Mann Whitney U test: W = 133,727, p < .01), see for mood score frequency.
Fig. 1 Frequency plots of mood scores via weekly SMS mobile phone messages for depression and mania in the stable group (left hand side) versus the unstable group (right hand side). QIDS-SR = Quick Inventory of Depressive Symptomatology – (more ...)
Mental imagery measures – bipolar patients versus healthy controls
Two sets of comparisons were conducted using the mental imagery scores, the first being between the bipolar patients and the matched healthy controls and the second being between the stable and unstable bipolar disorder patients.
General use of imagery (see , Panel A)
Significantly higher levels of general imagery use (SUIS) were reported by the patients with bipolar disorder compared to healthy controls, t(44) = 2.1, p = .042, Mdiff = 0.41, 95% CI [0.02, 0.80]. The Verbal versus Imagery Processing Style VAS (, Panel A) scores indicated higher imagery processing in the bipolar group t(44) = 2.82, p = .007, Mdiff = 1.43, 95% CI [0.41, 2.46], together with lower verbal processing t(44) = 2.58, p = .013, Mdiff = 1.35, 95% CI [0.29, 2.40].
Fig. 2 Results on a range of imagery measures for bipolar patients (black bars) versus non-clinical controls (white bars) (mean ± sem). SUIS = Spontaneous Use of Imagery Scale; VAS = Visual Analogue Scale; (more ...) Imagery of future events (see , Panel B)
When deliberately generating images of negative future scenarios, the bipolar group had higher vividness ratings (PIT Neg), t(44) = 2.1, p = .039, Mdiff = 0.65, 95% CI [0.03, 1.27] than the healthy controls. There was no significant difference between groups for the positive scenarios, t(44) = 0.48, p = .63.
The patients with bipolar disorder also reported higher levels of intrusive imagery of future events (IFES Total Score), t(44) = 2.17, p = .036, Mdiff = 11.42, 95% CI [0.79, 22.06]. A significantly higher proportion of negative events was reported by patients (0.47) than controls (0.24), t(44) = 2.8, p = .007, Mdiff = 0.23, 95% CI [0.07, 0.40].
Imagery-based interpretation bias (see , Panel C)
When confronted with ambiguous stimuli (HIT), compared to controls, the bipolar group reported significantly fewer positive homographs (t(44) = 2.72, p = .009, Mdiff = 1.17, 95% CI [0.30, 2.04]) and a statistical trend for more negative homographs (t(44) = 1.88, p = .067, Mdiff = 0.83, 95% CI [−0.06, 1.71]). However, when patients did demonstrate a positive bias, the ‘vividness’ of their positive images (see ‘Positive Vivid’ on , Panel C) was greater than for the controls, t(44) = 2.2, p = .032, Mdiff = 0.84, 95% CI [0.08, 1.60].
Mental imagery measures – bipolar patients with stable mood versus unstable mood
There were no significant differences between the stable and unstable groups in terms of general imagery use (SUIS or Verbal vs Imagery Processing Style VAS; t’s < 0.98, p’s > .34), deliberate generation of future events (PIT; t < 2.04, p > .05), or interpretation bias (HIT; t’s < 2.18, p’s > .04).
However, intrusive imagery of the future (IFES) was associated with greater mood instability. Compared to the stable patient group (M = 21.32, SD = 14.24), the unstable patient group (M = 50.68, SD = 17.3) had higher IFES Total Scores, t(21) = 4.4, p < .001, Mdiff = 29.36, 95% CI [15.43, 43.29]. This imagery consisted of a significantly higher proportion of negative events in the unstable group (0.71) than the stable group (0.21) t(21) = 4.8, p < .001, Mdiff = 0.5, 95% CI [0.28, 0.71]. Across the sample, there was no correlation between IFES Total Score and baseline mania score on the ASRM [r(21) = −.07, p = .76]. IFES Total Score was strongly associated with high levels of depression (BDI-SF) (r(21) = .87, p < .001, 95% CI [0.72, 0.95]), and with anxiety (STAI) (r(21) = .76, p < .001, 95% CI [0.51, 0.89]), see .
Positive correlation between prospective intrusive imagery and anxiety in bipolar patients. IFES = Impact of Future Event Scale; STAI = Spielberger State Anxiety Inventory.