Do Psychiatric Conditions Change the Way We See Faces and Emotions?
Review | Neuroimaging| Milly Darragh
This article reviews a meta-analysis regarding how neutral faces may not be neutral for axis-I psychiatric disorders. We explore the mechanisms of why each disorder may not see a neutral face the same way a neurotypical mind might.
A review of: Rethinking the use of ‘Neutral’ faces as a baseline in fMRI neuroimaging studies with Axis-I Psychiatric Disorders (Filkowski & Haas, 2016)
When you look at the following face, which emotion do you see?
Figure 1. An example of neutral face stimuli. Adapted from an image by STYLEPICS from depositphotos. https://depositphotos.com/13653328/stock-photo-handsome-caucasian-man-blue-eyes.htm.
Happy? Sad? Angry? Scared?
Statistically speaking, you probably said ‘neutral’ or ‘normal’. That is because this stimuli has been designed and used as a ‘neutral stimuli’ in psychology and neuroscience in many experiments and studies [1]. This face is supposed to represent no emotions, or act as a palette cleanser in experiments where multiple faces or emotions are being presented. However, does this stimulus actually work as planned? Questions regarding the subjectivity of a neutral face in psychological studies are growing, yet research diving into the effectiveness of this method has only just begun [1]. Specifically, how do different cognitive conditions affect the processing of a neutral face? With mental health conditions being one of the most common conditions worldwide, Filkowski & Haas (2016) explored why it may be time to rethink the use of neutral face stimuli in neuroscientific research [1].
Many psychological and neurological studies have been published using neutral face stimuli as an experimental technique. Facial stimuli in scientific experiments can be used to test a multitude of effects, with well-documented activity in the amygdala, the fusiform gyrus, and the prefrontal cortex of the brain [1]. Of course, we know that the brain cannot be separated into specific functions, yet these areas have continually shown activation when participants are exposed to facial stimuli. The use of neutral face stimuli assumes that each patient will respond and process these stimuli the same way in order to create a baseline of activity; however, this may not be the case [1].
The validity of the ‘neutral face’ across participants have recently been called into question – how can we ensure that each participant is processing these stimuli as intended? A large aspect of this issue lies in how neurodiverse brains process stimuli differently from neurotypical brains [1]. Whilst some research has investigated the effect autism, ADHD, or Tourettes have on facial processing,
little work has explored how mental illness influences this processing [1]. Axis I Disorders refer to a class of disorders that are considered mental health issues, and refer to the most common mental illnesses. These include anxiety disorders, mood disorders, substance-abuse disorders, and behaviour disorders [1]. Interestingly, the abnormal processing of emotional faces in patients with Axis-I disorders has been recorded across many neuroscience and psychological studies. Axis-I Disorders are fairly prevalent in the general population, estimated as affecting 1.6% - 12.2% of people. Therefore, cognitive differences attributed to Axis-I Disorders could have significant impacts on previous, and current studies.
Major Depressive Disorder (MDD) is one of the most common mental illness diagnoses, affecting 5% - 18% of the population. This disorder involves low mood, loss of interest in daily activities, severe fatigue, and other physiological symptoms [1]. Neurologically, MDD patients have shown changes in the frontal lobe, amygdala, hippocampus, temporal lobe, and thalamus. Furthermore, decreased levels of activity in serotonin circuits have been shown in MDD patients. These changes have been shown to significantly impact behaviour, mood, and cognition, so the application of baseline measures to neutral face stimuli is certainly important [1].
Bipolar Disorder (BD) is an Axis-I disorder that is characterised by periods of depressive and elated mood. This disorder can have significant impacts on a patient’s daily life and affects approximately 4.4% of the wider population. Prefrontal cortex abnormalities, as well as a clear link between the thinning of the cortex and episodes of mania, have been observed in BD using structural MRI [1]. Decreased volume in the frontal lobe of BD patients has also been detected, a section of the brain involved with future thinking and problem-solving.
Anxiety Disorders are a collection of mental illnesses that involve overthinking, panic, avoidance behaviours, and exaggerated worry. The most common types of anxiety disorders include GAD (Generalised Anxiety Disorder), OCD (Obsessive-Compulsive Disorder), Panic Disorder, SAD (Social Anxiety Disorder), and PTSD (Post-Traumatic Stress Disorder) [1]. These disorders can vary in their pathology, severity, and symptomatology, but impaired hippocampal and prefrontal cortex functioning have been observed across the varying types of anxiety.
Schizophrenia is a complicated disorder where patients struggle to differentiate and interact with reality and fiction. Schizophrenia has hereditary influences and affects approximately 0.3% of the population [1]. The main symptoms of this disorder include deluded thinking, hallucinations, and beliefs that interfere with a patient’s ability to function. One of the main differences of a schizophrenic brain is the decreased volume of temporal lobe matter [1]. This difference supposedly contributes to the auditory and visual hallucinations that patients with schizophrenia experience.
Knowing what we now know about mental illnesses, what impacts have these conditions had on neutral face stimuli?
Depression and Neutral Faces
Filkowski & Haas (2016) observed a negative bias associated with MDD and neutral face stimuli. This would suggest that decreased neural activity has been found in MDD patients, specifically in areas such as the amygdala, medial prefrontal cortex, and the dorso-lateral prefrontal cortex. Each of these regions of the brain has been associated with certain cognitive processes and roles in neuroscience. For example, the amygdala has shown significant activity in facial recognition studies, therefore, decreased activity in this area was almost expected [1]. Abnormal (decreased) activity leading to abnormal cognitive processing is a concept that thoroughly agrees with the current research regarding cognitive mechanisms. Since the amygdala is heavily associated with facial processing and emotional regulation, this would account for the changes seen in MDD patient data [1].
The medial Prefrontal Cortex (mPFC) is involved in memory recollection, specifically episodic memory. This type of memory uses the experiences of an individual to understand the context of events, memories, or information [1]. Neurotypical individuals in experiments have reported activity in the mPRC in response to both neutral faces and emotional faces, and a major difference in MDD patients was reported within this activity as well. Less activity in the mPFC of MDD patients is supposedly due to the decreased activity of serotonergic circuits in the brain that MDD patients experience [1].
Finally, the dorso-lateral prefrontal cortex (DLPFC) has also had decreased activity recorded in MDD patients upon neutral stimulation. This area of the brain is associated with executive control, meaning cognitive tasks such as attention, working memory, and planning [1]. The role of working memory is colloquially known as short-term memory, and therefore can be believed to play a role in facial recognition in addition to other cognitive tasks. Furthermore, studies have shown that successful antidepressant treatments of MDD patients increased the use of the DLPFC across a multitude of cognitive tasks.
One consistency found across a multitude of studies is the abnormal frontal lobe activity observed in MDD patients presented with neutral face stimuli. The frontal lobe is involved in many areas of cognition and cannot be separated into areas of specific function. Instead, we can infer that abnormal frontal lobe activity can explain the observed behaviours of MDD patients [1]. However, this data also showed variance of amygdala activity in MDD patients, with some studies showing increased activity but others reporting decreased activity. Amygdala activity has been altered to neurotypical levels when MDD patients have successful antidepressant treatments, which would certainly suggest the impacts of MDD are a main factor of the abnormal activity observed in the amygdala [1]. Whilst the literature may be conflicting regarding how MDD processing interacts with amygdala activity, abnormal processing is certainly present in this condition.
Bipolar & Neutral Faces
A clear negativity bias has been reported in patients with Bipolar Disorder (BD); specifically, BD patients identify neutral faces as sad or upset [1]. Previous literature discusses the abnormal activity of the amygdala, hippocampus, and prefrontal cortex (DLPFC and inferior frontal gyrus) BD patients [1]. The presence of abnormal amygdala cognition in BD patients is inconsistent regarding the mechanisms of processing. Some studies have reported decreased firing rates, but others have shown increased firing rates. However, decreased DLPFC activity is well documented in BD patients. This circuit involves executive function (such as attention, coordination, and inhibition), and a decrease in activity could account for cognitive disturbances that are experienced with BD [1]. Specifically, decreased DLPFC activity may contribute towards the planning and motivation inconsistencies that are observed in BD patients. Filkowski & Haas (2016) discuss the emotional attachment that BD patients show when presented with neutral face stimuli, even suggesting that the strong correlation of decreased DLPFC activity may potentially be used as a biomarker for individuals at risk or suspected BD [1].
Other areas of abnormal processing include the insula and anterior cingulate cortex (ACC). The ACC links the limbic system (emotional cognition) and the prefrontal cortex (practical cognition), meaning decreased ACC activity may be affecting both of these elements of healthy cognition [1]. Furthermore, the DLPFC is also involved in working memory and executive function. Again, decreased DLPFC activity may explain some of the abnormalities observed in the behaviour of BD patients and the changes that occur with processing neutral face stimuli. A unique concept in BD is how the differing mental states (manic vs. depressed) may affect the results of all studies regarding neutral face stimuli. Currently, research is unclear on how these two varying states impact the emotional and cognitive processing of stimuli, such as neutral vs. emotional faces.
Anxiety & Neutral Faces
Anxiety is a broad spectrum, with each subsection of anxiety affecting cognition and neural activity in different ways. One study has focussed on how anxiety disorders are overall affected in their processing – with the key finding that amygdala activity was increased when compared to healthy controls [1]. The amygdala is not only involved in facial processing, but is a hub for fearful and threatening stimuli processing and output. The increased activity would then suggest that patients with anxiety disorders are experiencing hyperactivity of fear centres in their brains [1]. Furthermore, studies using different stimuli found activation of the amygdala produces pharmacological anxiogenic (anxiety-inducing) effects, whilst inactivation of the amygdala produces pharmacological anxiolytic (anxiety-reducing) effects [1]. This concept can be applied to our understanding of how patients with anxiety process neutral face stimuli to understand the data we have on amygdala hyperactivity.
Social anxiety disorder (SAD) is a very common form of anxiety and manifests in the worries and fearfulness of social situations. SAD patients had observed increased amygdala activity, believed to be the result of patients unsuccessfully searching for social cues within the neutral stimulus, resulting in increased fear and anxiety [1]. SAD patients also reported higher insula and superior temporal sulcus (STS) activity compared to healthy controls. The insula has a well-established connection with anxiety disorders and behaviours, and anxiolytic effects (benzodiazepines, antidepressants, meditation etc.) have decreased firing rates in the insula. Aside from the link between the insula and anxiety, strong correlations between emotional processing and increased neural activity have been observed in this region [1]. This would support the idea that SAD increases the emotional processing of patients in an attempt to focus on non-existent social clues in neutral faces, which are regulated by anxiogenic regions of the brain such as the amygdala and insula. The STS is a region of the brain associated with theory of mind, or the ability to understand other’s emotional and mental processing [1]. An increased firing rate in this region would again suggest that SAD patients are searching for emotions and social cues in a neutral face, which is not a process observed with healthy controls.
Panic Disorder (PD) is a form of anxiety which presents as frequent, unexpected panic attacks when no threatening stimuli are present to elicit this response. PD patients have shown clear increases in amygdala activities, suggested to be due to the increased cortisol levels, fear signals, and perception of threatening stimuli. Aside from the amygdala, the anterior cingulate cortex (ACC) has reported abnormal processing in PD patients [1]. This area of the brain is involved in processes such as emotional regulation, attention, and mood regulation. Increased firing and activity in the ACC may be occurring in PD due to the hypervigilance PD patients show towards their environment.
Interestingly, another study found that healthy controls who displayed anxiety tendencies had longer periods of amygdala processing when compared to healthy controls who did not display anxiety tendencies. This would again suggest that the amygdala plays a large role in anxiety disorders and behaviours.
Overall, this data supports the idea that abnormal processing of neutral facial stimuli occurs within patients with anxiety disorders – and anxiety tendencies. The main regions of interest for abnormal processing in anxiety disorders is the limbic system which includes the amygdala, ACC, and insula.
Schizophrenia & Neutral Faces
There is quite a consistent difference between healthy controls and Schizophrenic patients’ processing of neutral faces, with schizophrenic patients attaching emotions to the neutral faces presented to them [1]. The association that schizophrenic patients make with emotions and neutral faces varies between studies and between patients, whether it is happy, sad, angry, or scared emotions observed — but neutral faces are usually not seen as neutral. Schizophrenics experience delusions, hallucinations, and struggle to define reality, so it is much easier for clear emotions to be seen in a face that is designed to be emotionless [1]. This is often a large issue with schizophrenia patients, as the inability to correctly identify emotions can lead to miscommunications, social issues, and amplified delusions.
Filkowski & Haas (2016) found consistent evidence that hypoactivity of the ACC may be a contributing factor to the abnormal emotional processing that schizophrenic patients experience. Decreased grey matter in the ACC has been a consistent finding in MRI studies of schizophrenic patients, with this atrophy preceding schizophrenia onset, suggesting a potential biomarker for at-risk schizophrenic patients [1].
Similarly, the DLPFC is associated with dysfunction in schizophrenic patients, accounting for some of the behaviours observed in schizophrenic patients, such as abstract concept regulation, using appropriate responses, working memory, and attention [1]. Reduced engagement with the DLPFC in these patients may explain why reduced ability of these executive functions is observed in schizophrenic patients.
There is a significant increase in amygdala activity that is observed in schizophrenic patients, as well as increased amygdala activity in emotional face stimuli when compared to healthy controls. Since the amygdala has a strong relationship with emotional processing and output, increased activity with schizophrenics may contribute strongly to the observed behaviour of hyper emotional attachment and an inability to accurately process emotionless stimuli [1]. As well as the limbic system experiencing abnormal processing, the hippocampus — a region of the brain largely dedicated to memory — has shown increased activity in schizophrenics. However, some studies have reported a reduced rate of firing in the hippocampus of schizophrenics, but Filkowski & Haas (2016) chalk this up to differences in study design and sample size. Varying schools of thought have attempted to explain why increased or decreased hippocampal activity is observed, yet neuroscientists will agree that abnormal activity in the hippocampus is not unexpected for schizophrenia patients.
24 out of 26 studies analysed by Filkowski & Haas (2016) showed significant differences between Axis-I disorders and healthy controls when shown neutral face stimuli. Specifically, these studies showed that frontolimbic and prefrontal areas (amygdala, ACC, DLPFC). Clearly, neutral faces are not perceived as such by patients with Axis-I disorders, showing that this stimulus may not be a robust baseline measure in neuroimaging studies [1]. Furthermore, some studies have shown that the choice of neutral face stimuli can alter results in healthy controls, suggesting that this choice of a baseline measurement may not be as accurate as neuroscientists think.
One of the largest things to take into consideration with this article is the practicality of this knowledge. Many studies involved here do not disclose or discuss the medications or treatments that patients are undergoing for their diagnosed conditions. For example, a patient with MDD who is undergoing successful treatment via antidepressants and psychotherapy may have shown very different results if they did not have these interventions [1]. The effectiveness of mental health treatments is still being discovered, with many inconsistencies found within treatment options. These demographics are important to consider but are certainly easier said than done. Clearly, this paper presents evidence that suggests directional abnormal activities across axis-I disorders, as well as the observation that certain regions of the brain are affected by these illnesses in relation to neutral face stimuli.
A specific issue with this topic is the nature of mental illnesses and the unpredictability that follows. For example, a BD patient in mania could strongly differ from the same patient in a depressive episode. This idea supports how such common and affective mental disorders are overlooked in academia, even within the subjects of psychology and neuroscience [1]. This article suggests that previous, current, and future research regarding any of the topics covered may be brushing past a large demographic factor — how do axis-I disorders impact processes that are considered neutral or normal for neurotypical participants? Approximately 1/5 people live with a mental illness worldwide, and this paper suggests that academia may have been using an inaccurate baseline measurement tool across a multitude of cognition and behavioural studies.
Milly Darragh - BAdvSci (Hons), Cognitive Neuroscience
Milly is entering her 4th year of the Bachelor of Advanced Sciences (Honours) programme in cognitive neuroscience. As an honours student she is fascinated by translational neurology and neuroscience, specifically neurodegenerative diseases. She is the current vice-president of scientific, and the president of UoA Campus Neuroscience Society.