Transcranial magnetic stimulation demonstrates a role for the ventrolateral prefrontal cortex in emotion perception

Highlights

• vlPFC is linked to emotion processing, but causal role has not been tested.

• We used TMS to disrupt vlPFC activation while showing angry and happy faces.

• Perceptual sensitivity increased for angry faces but decreased for happy faces.

• Results demonstrate causal role for vlPFC in emotion perception.

• Findings have implications for noninvasive treatments of mood disorders.

Abstract

The lateral prefrontal cortex, a region with both structural and functional connectivity to the amygdala, has been consistently implicated in the downregulation of subcortical-generated emotional responses. Although previous work has demonstrated that the ventral lateral prefrontal cortex (vlPFC) is important to emotion processing, no study has interrupted vlPFC function in order to test is role in emotion perception. In the current study, we acutely disrupted vlPFC function in twenty healthy adult participants by administering sham stimulation and transcranial magnetic stimulation (TMS), in randomized order, during performance of an emotional perception task. During sham stimulation, participants demonstrated increased perceptual sensitivity for happy faces compared to angry faces. Disruption of the vlPFC eliminated this difference: in this condition, perceptual sensitivity did not differ between happy and angry faces. Reaction times and response bias did not differ between emotions or TMS conditions. This pattern of perceptual bias is consistent with effects observed in a wide range of affective disorders, in which vlPFC dysfunction has also been reported. This study provides insight into a possible mechanism through which the vlPFC may contribute to emotion perception.

Introduction

Biased emotion perception is characteristic of individuals with mood disorders, including anxiety (Stein et al., 2007), depression (Disner et al., 2011), and post-traumatic stress disorder (PTSD) (Felmingham et al., 2014). Aberrant emotional processing is also observed in individuals with schizophrenia and bipolar disorder (Delvecchio, et al., 2013, Hill, et al., 2008). In particular, people with mood and other psychiatric disorders frequently display enhanced sensitivity to negatively valenced emotional stimuli (Monk et al., 2006), with some disorders (or subsets of disorders) also displaying decreased sensitivity to positively valenced stimuli (i.e., a positive blockade; (Disner et al., 2011, Keding and Herringa, 2016, Kerestes et al., 2016, Schechner et al., 2012, Winer and Salem, 2016). This aberrant emotional processing is thought to result from dysregulated connectivity between higher-order control regions in the prefrontal cortex and subcortical regions such as the amygdala (Disner et al., 2011, Dixon et al., 2017, Motzkin et al., 2015, Ochsner and Gross, 2005, Quirk and Beer, 2006, Salzman and Fusi, 2010, Shin et al., 2006).

Transcranial magnetic stimulation (TMS) involves administering noninvasive magnetic pulses that temporarily disrupt or enhance the function of the target region, depending on the frequency of pulses (Hallett, 2007). As such, TMS is a promising tool for testing hypotheses about the role of targeted brain regions in cognitive and emotional processes (Schwarzkopf et al., 2011). By disrupting the function of a prefrontal cortical region with known connectivity to the amygdala, and observing effects on behavior, it is possible to assess whether the target region plays a role in that behavior (Hartwigsen et al., 2015). Although lesion work suggests that the medial prefrontal cortex is important to emotion regulation (Heidbreder and Groenewegen, 2003), this region is inaccessible to noninvasive stimulation techniques such as TMS. Within the lateral prefrontal cortex, anatomical tract-tracing work in non-human primates has identified anatomical connections between the ventrolateral prefrontal cortex (vlPFC) and the amygdala, which occur through largely uni-directional projections to the basolateral amygdala (Ghashghaei et al., 2007). Since the vlPFC lacks a readily apparent rodent homolog (Shiba et al., 2016) but is accessible to noninvasive stimulation in humans, its investigation in humans and non-human primates is particularly important (Kim et al., 2018, Marques et al., 2018, Shiba et al., 2016).

The vlPFC has been implicated in a broad range of cognitive functions, including response inhibition (Hampshire et al., 2010), working memory (Yaple and Arsalidou, 2018), priming (Yaple and Arsalidou, 2017), attention (Arsalidou et al., 2013), and social interactions (He et al., 2018, Liu et al., 2016). Among these functions, vlPFC activation has been reliably observed in response to emotional faces (Liu et al., 2016), according to meta-analyses of neuroimaging studies (Fusar-Poli et al., 2009, Zinchenko et al., 2018). The lateral prefrontal cortex (lPFC) is consistently implicated in processing and regulating basic emotions generated within subcortical regions, such as the amygdala (Dixon et al., 2017). As a putative mechanism for the influence of lPFC on emotion processing, anatomical tract-tracing work in non-human primates has identified anatomical connections between the ventrolateral prefrontal cortex (vlPFC) and the amygdala (Amaral and Price, 1984, Ghashghaei et al., 2007, Van Hoesen et al., 1972), and resting-state and task-based fMRI studies in humans have found the vlPFC and basolateral amygdala to be functionally connected (Etkin et al., 2009, Kerestes et al., 2017, Roy et al., 2009). If disrupting the lateral PFC were to alter emotional behavior, it would not necessarily follow that these behavioral effects resulted from interrupted PFC regulation of the amygdala, as the lateral PFC also shares functional and structural connections with other cortical and subcortical regions (Leh et al., 2007). However, such a finding would generate hypotheses for future research, with potential implications for noninvasive treatment (Kim et al., 2018, Perera et al., 2016).

Activity in the vlPFC tracks perceived emotion even under near-threshold perceptual conditions (György Szabó et al., 2017, Kohno et al., 2015, Pessoa et al., 2006, Pessoa and Padmala, 2005, Wang et al., 2014). Moreover, aberrant vlPFC activation to emotional faces is observed across a broad spectrum of mood disorders, including generalized anxiety disorder (Holzel et al., 2013), major depressive disorder (Henderson et al., 2014, Kerestes et al., 2016), bipolar depression (Hafeman et al., 2014, Passarotti et al., 2011, Phillips and Swartz, 2014), and posttraumatic stress disorder (Keding and Herringa, 2016). However, to date, no study has directly manipulated the vlPFC while studying emotion perception in humans. We assessed whether the vlPFC plays a role in emotion perception by acutely disrupting its function in healthy individuals during performance of an emotion perception task. Unlike when single TMS pulses are delivered at rest during TMS/fMRI, delivery of TMS pulses during execution of a task interacts with ongoing neural activity, frequently disrupting endogenous patterns of brain activity important for task performance (here termed “task-directed TMS;” (Hartwigsen et al., 2015, Rose et al., 2016, Schwarzkopf et al., 2011). lPFC abnormalities have been linked to symptomatology of a range of affective disorders (for review see Dixon et al. (2017)). vlPFC activation has been associated with blunted emotional experience, emotional downregulation, and reduced amygdala activation (Adolphs, 2008, Breiter et al., 1996, Hariri et al., 2000, Kohno et al., 2015, Lieberman et al., 2007). Similarly, vlPFC lesions in non-human primates increase anxiety behavior, and vlPFC inactivation results in avoidance of negative emotional stimuli (Agustin-Pavon et al., 2012, Clarke et al., 2015). However, other studies suggest the opposite direction of effects (Blair et al., 1999, Dal Monte et al., 2014).

In addition to assessing a putative role for the vlPFC in emotion perception, a secondary goal of this study was to test whether the vlPFC affects perception of both negatively and positively valenced stimuli, and, if so, whether the effects are in opposite directions. Studies comparing negative to neutral stimuli have consistently found a bias for enhanced perception of negatively valenced (i.e., threat-related) stimuli (Agustin-Pavon et al., 2012, Chang et al., 2017, Kohno et al., 2015). Studies comparing positive to neutral stimuli have generally found a bias away from positively valenced stimuli, though the pattern is less consistent than for negative valence (Feldmann‐Wüstefeld et al., 2011, Frewen et al., 2008, György Szabó et al., 2017, Schechner et al., 2012, Winer and Salem, 2016). For example, healthy youth displayed higher amygdala-vlPFC connectivity when viewing angry faces than when viewing happy faces, but youth with PTSD displayed the opposite pattern (Keding and Herringa, 2016). Based on these previous findings, we predicted that using TMS to disrupt vlPFC activation in healthy individuals would result in a shift toward behavioral patterns observed in mood-disordered individuals: exaggerated perception of threat, and diminished perception of positively valenced stimuli.