Bipolar disorder (BD) is a severe and chronic psychiatric condition associated with functional impairment and disability (Fagiolini et al.
2005; Michalak et al.
2007). Recent models of BD stress the importance of difficulties in positive emotion processing (Gruber
2011; Phillips and Vieta
2007). An important next step is to characterize processes that may underlie and contribute to these patterns of emotion disturbance. One promising route is to explore attentional biases, which have been shown to play an important role in depression (e.g., Gotlib et al.
2004; Joormann et al.
2007). However, comparably less is known about how attentional biases contribute to disturbances in positive emotion characteristic of BD.
Emotion disturbance in BD
Recent models suggest that individuals with BD experience heightened and persistent elevations in positive emotionality (e.g., Gruber
2011). This pattern is consistent with psychosocial models of BD that implicate heightened reward seeking and goal-striving (e.g., Alloy et al.
2009; Johnson
2005). For example, individuals at risk for BD and remitted BD adults self-report greater positive affect than healthy controls in response to emotional films (Gruber et al.
2011a,
b), photos (M’Bailara et al.
2009), and at the prospect of earning rewards (Meyer et al.
2001). BD individuals and individuals at risk for BD also exhibit increased parasympathetic reactivity in response to emotional stimuli such as films, photos, and memories (Gruber et al.
2008,
2009; Sutton and Johnson
2002). Neuroimaging studies reveal that people with BD exhibit increased activity in brain regions typically associated with reward (e.g., ventral striatum) to positive stimuli (e.g., Dutra et al.
2015; Wessa et al.
2007). Heightened positive emotionality differentiates BD from major depressive disorder (Gruber et al.
2011a), and has important implications for psychosocial treatments aimed at reducing positive emotionality and subsequent manic episodes (e.g., Johnson
2005).
Abnormalities in negative emotionality might also be expected in BD, given the characteristic frequent and recurrent episodes of depression (Judd et al.
2003). However, the current literature suggests that people diagnosed with and at risk for BD may not differ from healthy controls in their emotional responses to negative stimuli including negative social feedback (e.g., Ruggero and Johnson
2006) and interpersonal criticism (Cuellar et al.
2009). At the same time, individuals with BD do report increased tendencies toward behavioral inhibition and neuroticism which is associated with increased negative affect (e.g., Alloy et al.
2008; Meyer et al.
2001; Murray et al.
2007).
Coinciding cognitive and emotional processes
Though many studies have evaluated abnormal emotional reactivity in a variety of populations, a critical next step includes isolating the correspondingly impaired cognitive processes in BD (Gruber
2011). Cognitive processes such as attention have long been recognized as contributing to, and being affected by, emotion (e.g. Schwarz
2000; Shimojo et al.
2003), and visual attention in particular has been shown to impact emotional responding (e.g., Cacioppo et al.
2000). One methodology well-suited to examine visual attention for emotional events is eye-tracking technology, which has allowed researchers to elucidate different patterns of visual attention to negative stimuli in mood disorders such as depression and BD (Gotlib and Joormann
2010; Gotlib et al.
2004; Mathews and MacLeod
2005).
Research suggests that adults diagnosed with depression take longer to disengage attention away from sad faces, which has been predictive of sustained negative mood (Kellough et al.
2008; Sanchez et al.
2013). A meta-analytic review of eye-tracking studies has also found that failing to attend to positive events was associated with reductions in pleasure in depressed individuals (Armstrong and Olatunji
2012). Research on BD has found impaired recognition of negative facial expressions, failure to demonstrate attentional biases towards negative stimuli (Elliott et al.
2000; Lembke and Ketter
2002), and difficulties maintaining negative emotional information (Gruber et al.
2013). These findings suggest that BD may be associated with attention
away from negative stimuli, while depression is associated with attention
toward it.
There has been comparatively less work examining attentional biases underlying
positive emotion. Wadlinger and Isaacowitz (
2008) found that after experimentally training healthy adult participants to selectively attend to positive information, they spent significantly less time looking at negative stimuli following the attentional training. Similarly, trait happiness is associated with increased attentional bias toward a variety of positive stimuli (Raila et al.
2015), and optimism is associated with gaze preferences away from negative health-related images (e.g., cancer tumor images; Isaacowitz
2005). Furthermore, older adults, who report higher and more stable levels of positive emotional states relative to younger adults (for review, see Lohani et al.
2013), display attentional preferences toward positive and away from negative faces (Isaacowitz et al.
2006b,
2008) and negative images (Noh et al.
2011).
However, studies on visual attention to emotional stimuli in BD have suggested that biases may be mood-congruent. For example, research has found that mildly depressed BD participants demonstrate bias towards negative and away from positive words using a modified dot-probe task compared with controls (Jongen et al.
2007). Other studies have utilized “free-looking” tasks to simultaneously display several images to subjects have demonstrated that currently manic BD adults attend more to positive images, and currently depressed BD adults attend less to positive images, as compared to controls (García-Blanco et al.
2013,
2014,
2015,
2017; Leyman et al.
2009).
These recent studies have suggested that selective attention may coincide with positive and negative mood states in BD, and that these attentional biases are not present during periods of remission. This theory has been bolstered by studies that report no group differences between those with remitted BD and healthy controls in attention bias for positive or negative faces using emotional dot-probe, and free-view, tasks (e.g., Peckham et al.
2015,
2016). Thus far, a substantial body of literature suggests that attentional processes may not harmoniously fit into the model of BD as a disorder of positive emotionality throughout all mood phases of the disorder. However, findings thus far have only investigated attentional biases to differently-valenced, simultaneously presented affective stimuli.
The present investigation
The present investigation aimed to explore attentional biases toward emotionally relevant stimuli in BD by employing continuous eye-tracking during individually presented images. By presenting images one at a time, the current study constrained participants’ potential attentional biases to affectively salient vs. non-salient areas, mimicking visual field processing in a more ecologically valid way. Using this methodology, individual and age-group differences in attentional biases have been found within non-clinical populations (Isaacowitz and Choi
2011; Wadlinger and Isaacowitz
2008). For instance, compared to younger adults, older adults have been found to focus their visual attention away from more negative regions and towards non-affective regions of an image (Noh et al.
2011). Using this method, we investigated whether individuals with remitted BD would deploy attention toward more salient negative or positive regions, as opposed to the non-emotional regions, of images, and whether or not these patterns of gaze fixation would differ from healthy controls. Considering the positive emotion persistence observed in remitted BD, two hypotheses were formed (Gruber
2011). The first predicted that individuals with BD would demonstrate specific attention biases towards positive emotional stimuli (i.e.
positive amplification). The other non-mutually exclusive hypothesis predicted that individuals with BD would demonstrate specific attentional biases away from negative emotional stimuli (i.e.
negative attenuation).’
We also conducted post hoc exploratory analyses to investigate early, relative to late, phases of image viewing as well as sustained attention across the entire duration of image viewing. This approach has shown promise in other domains of emotion and attention processing such that individuals who report higher trait levels of happiness have shown increased dwell time and fixation counts to positive compared to neutral stimuli in later, rather than early, stages of visual attention (e.g., Nakayama and Mackeben
1989; Raila et al.
2015).