Behavioral and electctrophysiological evidence for enhanced sensitivity to subtle variations of pain expressions of same-race than other-race faces

Highlights

• Study neural responses to subtle variations of pain expressions.

• Measure EEG responses to increasing amount of pixels of pain expressions.

• P2 amplitude increases to the increasing amount of painful pixels in faces.

• P2 amplitude changes is more sensitive to same-race than other-race faces.

Abstract

Recent research has shown enhanced brain activity in responses to pain expression of same-race than other-race faces. However, due to the stimuli used in electctrophysiological (EEG) research (i.e., photos of static pain expressions) and low temporal resolutions of neural activity reported in functional magnetic resonance imaging (MRI) studies, it remains unclear how dynamic neural responses to subtle variations of pain expressions are shaped by perceived race in the first few hundreds of milliseconds after face onset. We tested the hypothesis of greater sensitivity to subtle variations of pain expressions of same-race than other-race faces by integrating psychophysical and EEG responses to increasing amount of pixels of pain expressions applied to neutral faces. Experiment 1 found that adults required fewer painful pixels to perceive pain expression of same-race than other-race faces. Experiment 2 showed that the amplitude of a positivity activity at 140–190 ms over the frontal/central electrodes (P2) increased faster in response to the increasing amount of painful pixels of same-race than other-race faces. These results provide evidence for enhanced sensitivity of both subjective feelings and neural responses pertaining to subtle variations of pain expressions of same-race than other-race faces.

Introduction

It is now generally agreed that empathy — psychological and neural processes underlying understanding and sharing of others' emotional states — plays a key role in prosocial behavior (De Waal, 2008; Decety et al., 2015). However, empathy for others' emotional states such as pain is strongly affected by interracial relationships between perceivers and targets. For example, White students reported greater feelings of empathy for White than Black's suffering and tended to give more help to White than Black individuals when making decisions regarding punishment and pain treatment (Johnson et al., 2002; Drwecki et al., 2011). Given the potential link between racial ingroup bias in empathy (RIBE) and racial bias in social behavior, recent brain imaging studies have investigated the neural mechanisms underlying RIBE extensively (Han, 2018; Molenberghs and Louis, 2018).

Early functional magnetic resonance imaging (fMRI) research on RIBE scanned Chinese and White (Caucasian) college students while they viewed video clips of faces with a neutral expression that received either painful or non-painful stimulation (Xu et al., 2009). It was found that the activities in the anterior cingulate (ACC) and supplementary motor area (SMA) were decreased in response to painful stimulations applied to other-race than same-race models. Other fMRI studies reported similar racial ingroup bias in empathic neural responses in the ACC, AI and somatosensory cortex (Contreras-Huerta et al., 2013; Azevedo et al.,2012; Shen et al., 2018) and in the temporoparietal junction (TPJ) and mediaprefrontal cortex (mPFC) (Mathur et al., 2010; Fourie et al., 2017). These findings suggest wide modulations of sensory, affective, and cognitive components of empathy (Shamay-Tsoory, 2010) by interracial relationships. However, fMRI studies cannot clarify the dynamic neural responses in the first few hundred milliseconds after stimulus onset to painful stimulation applied to others due to the low temporal resolution of blood oxygen level dependent (BOLD) signals.

The time course of RIBE has been investigated by recording electroencephalography (EEG) when viewing static photos of others in pain. An early EEG study analyzed event related potentials (ERPs) in response to photos of same-race and other-face faces with painful or neutral expressions (Sheng and Han, 2012). It was found that painful (vs. neutral) expressions increased the amplitude of a positive wave at 128–188 ms after stimulus onset (P2) and also induced positive shift of a negative wave at 200–300 ms (N2) over the frontal/central brain regions. Moreover, these effects were significantly stronger for same-race than other-race faces. Racial ingroup bias in these early neural responses to perceived pain was replicated in other studies that used either photos of pain/neutral expressions (Han et al., 2016; Sheng et al., 2016a, Sheng et al., 2016b, 2017) or photos painful stimulation applied to same-race and other-race faces (Sessa et al., 2014; Contreras-Huerta et al., 2014). While these ERP findings have revealed racial ingroup bias in fast neural responses to perceived pain in others, only static images of either pain expressions or painful stimulation were used. In real life situations, painful facial expressions usually change dynamically (e.g., from neutral to painful) rather than being perceived in a stationary state. Static images of faces or painful stimulation decontextualize sensorimotor information (Zaki and Ochsner, 2012) and thus do not allow researchers to probe how subjective feelings and neural activities vary in accordance with subtle variations of pain expressions and whether the psychological and neural responses to subtle variations of pain expressions are influenced by perceived race.

Previous research has suggested multiple-level responses to perceived facial expressions. For example, while viewing dynamic facial expressions of fear and happiness activated the occipital and temporal cortices (e.g., Sato et al., 2015), disrupting the sensorimotor responses using repetitive transcranial magnetic stimulation led to impaired recognition of these facial expressions (Pitcher et al., 2008). These findings suggest that viewing others' emotional expressions is associated with not only enhanced perceptual processing but also motor production of the perceived facial expression in observers and such sensorimotor simulation triggers neural responses involved in experiencing the corresponding emotion (Goldman and Sripada, 2005; Wood et al., 2016). Viewing others' painful expressions might engage sensorimotor simulation either because an early frontal response to others' pain vs. neutral expressions was reduced when observers' facial muscle movement was prevented (Han et al., 2016). Moreover, viewing others' painful expressions also activated brain regions involved in conflict monitoring (e.g., ACC) and affective sharing (e.g., AI) (Saarela et al., 2006; Sheng et al., 2014). Similar to the modulation of empathic neural responses by perceived race (Han, 2018), neural activities underlying sensorimotor simulation are also modulated by perceived race (Losin et al., 2012; Earls et al., 2013). Moreover, the effect of inhibition of facial muscle movement on the neutral response to pain expressions was evident when viewing same-race but not other-race faces (Han et al., 2016). Taken together, these findings allowed us to postulate stronger modulations of subjective and neural responses to subtle variations of pain expression of other-race than same-race faces.

The current work tested the hypothesis that healthy adults are more sensitive to variations of pain expressions of same-race than other-race faces by integrating psychophysical and EEG measures. Because static images provides exact times of stimulus onset and fit the requirement of recording and analyzing ERPs, we used static pictures of morphed faces produced by integrating pixels from photos of faces with neutral and pain expressions. Morphed faces have been used in previous brain imaging studies to examine neural processes involved in face recognition (e.g., Uddin et al., 2005, Ma and Han, 2011) and emotion recognition (e.g., Campanella et al., 2002; Frenkel and Bar-Haim, 2011). In the present work, the increased amount of painful pixels in the stimuli represented subtle variations of pain expressions and allowed us to examine changes of subjective report of perceived pain of others and related brain activities in response to faces with the increased amount of painful cues. In Experiment 1 participants performed a forced choice task on morphed faces with increased amount of pixels from pain expressions. A psychometric Weibull function was constructed to estimate each participant's judgment performance. We predicted that fewer pain-pixels are required to generate perception of pain expression of same-race than other-race faces and that the frequency of perceiving pain expressions increases faster along the increased amount of painful pixels for same-race than other-race faces. In Experiment 2 participants performed a forced choice task on occasional novel target faces with pain or neutral expressions while EEG was recorded. We focused on the P2 amplitude to non-target morphed faces and tested the prediction that the P2 amplitude increased more rapidly as a function of the increased amount of painful pixels in morphed same-race than other-race faces. Previous questionnaire and brain imaging research suggest gender difference in empathy ability and neural responses to perceived facial expressions (e.g., Eisenberg and Lennon, 1983; Schulte-Rüther et al., 2008; Derntl et al., 2010). In addition, there was evidence that empathic neural responses are modulated by task demands in different ways in female and male adults (Han et al., 2008). Therefore, the present study only tested female participants to avoid potential confounds of gender difference in empathy.