Alternative explanations through the lens of neuroscience and predictive coding theory
Perception helps make sense of our world, translating information from our senses into lived, real experiences (Goldstein, 2018). But what happens when someone ‘perceives’ things that aren’t real? Auditory Hallucinations (AH)—hearing things that aren’t real—can be highly distressing, and one of the most perplexing phenomena of perception (Hugdahl & Sommer, 2018). Although primarily associated with schizophrenia, AH occur in multiple other contexts (Hugdahl, 2017), including Parkinson’s disease, epilepsy, anxiety, depression, and substance-abuse (Sommer, Koops, & Blom, 2012); in non-clinical populations (Badcock & Hugdahl, 2012); and in children (Mertin & O’Brien, 2013). Prevalence estimates vary from 5% to 20% (Hugdahl, 2017); however, significant stigma exists, and there is evidence of under-reporting and lack of help-seeking (Water, et al, 2018). For those who do seek help, one third to half do not respond to standard treatment of antipsychotic medications (Sommer, et al., 2012; Bobes, et al., 2003). It is important therefore to look more widely for explanations of AH. This paper explores AH in stress-related disorders such as anxiety, depression, and Post-Traumatic Stress Disorder (PTSD). Applying concepts from the neuroscience paradigm and predictive coding’s theory of perception, it offers alternative explanations for how AH could arise outside of a psychotic context, and potential alternative therapies.
Despite widespread prevalence of AH—around 15% in anxiety and PTSD and up to 40% in depression (Waters, et al., 2018)—few studies have investigated the underlying mechanisms outside of a psychotic explanation (Toh, Thomas, & Rossell, 2015). Known diatheses include trait anxiety (Hoskin, Hunter, & Woodruff, 2014), hypervigilance to auditory threat (Dudle, et al., 2014), and emotional and expressive suppression (Prochwicz, Klosowska, & Sznajder, 2018; Badcock, Paulik, & Maybery, 2011). Although accompanied by significant distress and impairment, it is not always accompanied by symptoms typical of psychosis (Waters et al., 2018). Yet the sole presence of AH, with no other features, permits a diagnosis of Schizophrenia under the DSM-5 (APA, 2013, p.122): indeed, excluding substance-abuse, this is the only diagnosis available (Waters, Blom, Jardri, Hugdahl, & Sommer, 2018)
The neuroscience paradigm approaches psychopathology as anomalies within the brain and nervous system (Altman, 2018). For example, depression and anxiety are hypothesised to result from deficits or excesses of neurotransmitters such as serotonin, dopamine, and norepinephrine (Kring, 2018); and/or from over-or-underactivity in certain brain areas, for example elevated activity in the amygdala and anterior cingulate, and diminished activity in the hippocampus, prefrontal cortex, and striatum (Kring, 2018; Fitzgerald, et al., 2017). PTSD patients also tend to have a smaller anterior cingulate (Baldaçara, et al., 2014), smaller amygdala and hippocampus (Logue, et al., 2018); and irregularities in hippocampus and amygdala shape (Akiki, et al., 2017). The neuroscience paradigm suggests that such anomalies interact in complex ways to affect mood and mental functioning.
In engaging neuroscience to explain AH, we frequently find increased activity and cortical reduction in the superior temporal gyrus (STG) of the temporal lobe, which contains the auditory cortex responsible for processing sounds; and a general pattern of reduced neural connectivity/plasticity in grey matter, and cortical thinning in white matter. This is particularly between different parts of the limbic system, associated with regulating emotion, and frontal lobes, associated with executive decision-making. There is also increased activity in the thalamus, the brain’s sensory relay system, and right striatum, associated with glutamates and dopamine. Neurotransmitter involvement in AH is less well understood (Bohlken, Hugdahl, & Sommer, 2017). Dopamine dysfunction is classically implicated, but evidence is accumulating that AH does not always involve abnormal dopamine function, raising questions about standard treatment using antipsychotics (Waters, et al., 2018). More likely factors may be increased glutamates and decreased GABA (Hugdahl, et al., 2015). Interestingly, neither serotonin nor norepinephrine—both strongly associated with anxiety and depression—appear to play any role in AH.
It is hypothesised that AH arises because of hyperactivation in bottom-up processing areas (the STG and thalamus) and hypoactivation in top-down processing areas (frontal lobes and general reduction in neural connectivity and cortical volume). Spontaneous bottom-up hyperactivation is not inhibited, because of top-down hypoactivation (Bohlken, et al., 2017; Hugdahl, et al., 2015).
To address this, therapy could focus on two key issues. Firstly, improving plasticity and cortical thickness/volume to redress top-down processing hypoactivity. Agomelatine is a new class of antidepressant that mimics the action of melatonin. Not an SSRI, it has successfully treated depression and anxiety—and has a unique ability to enhance neuroplasticity and neurogenesis in brain areas such as the hippocampus and prefrontal cortex (Jeon & Kim, 2016; De Berardis, et al., 2013). Augmenting or replacing SSRI therapy with agomelatine might improve neural networks and white matter connectivity, enabling better top-down processing responses to faulty sensory signals. The second issue, hyperactivity, is influenced by excess glutamate levels and/or glutamate receptors, which are normally balanced by GABA inhibitory influences (Hugdahl & Sommer, 2018). Pharmacological treatments could be provided to increase GABA and/or decrease glutamate activity.
One problem with the neurological paradigm is that, although accounting for AH in terms of anomalies in the brain, it does not explain causes. Predictive Coding Theory (PCT) helps fill this gap. PCT is a theory of perception that falls within the Predictive Brain paradigm broadly attributed to Andy Clark (2013). PCT suggests normal perception consists of interactions between top-down processing (knowledge) and bottom-up processing (stimuli-driven): neural networks predict expected input (top-down) and match these predictions—called prior expectations—to received stimuli (bottom-up). For example, seeing a vehicle approach predicts the sound of a vehicle approaching (Nazimek, Hunter, Hoskin, Wilkinson, & Woodruff, 2013). If there is a difference between the prior expectation and the stimuli received—if the external input does not match the expectation—an error signal, called the “prediction-error” is sent back to higher-up areas for subsequent processing (Nazimek, Hunter, & Woodruff, 2012). Prediction-error is not an “error” in functioning, but part of the prediction-matching process that helps make sense of our environment.
Prior expectations exist as knowledge and therefore as established neural pathways. “Priming” occurs in response to a prediction, such that target cells for the prior expectation increase firing in readiness. Frequently primed expectations, whether internal or external, are thought to have increased representational size (Nazimek, et al., 2012), and such well-worn neural pathways are hypothesised to play a crucial role in AH.
PCT suggests AH occurs as a failure of the prediction-error response, thereby permitting perception of a sound not validated by external stimuli. Friston (2005) found that normal perceptual inference must have a degree of in-built uncertainty, and that failure to encode this uncertainty gives undue influence to the prior expectation, leading to a false inference (hallucination). Building on this, Nazimek and colleagues (2012, 2013) offer the expectation-perception model of AH. They note that perceptions can be generated without external sensory input, and propose that AH can be driven by mental states; formed in the process of predictive coding; and translated into false perceptions because of prior expectations. That is, an anticipated perception activating a well-worn neural pathway can override prediction-error, manifesting in an acoustic experience. In their view, subjects perceive AH as having an external source because they have the acoustic qualities of real percepts.
Moreover, psychological stress and “trait anxiety” have a significant influence on AH and prediction-error failure in both clinical and healthy populations. This suggests that, as stress represents a response to perceived threat, in states of heightened anxiety it may be adaptive for perceptual systems to bias towards false positives (hallucinations), in preference to false negatives (threat or harm) (Hoskin, Hunter, & Woodruff, 2014).
In anxiety and depression, this makes compelling sense. In a state of high anxiety, well-worn neural pathways of expectation, activated perhaps through hypervigilance (Mertin, et al., 2013), rumination (Badcock, et al., 2011) or cognitive biases, can outweigh normal predictive coding and be experienced as genuine acoustic events, and this can be explained as adaptive. This approach explains, through a propensity to stress and trait anxiety, how AH can arise in apparently healthy populations. This perspective also accounts for the poor response to antipsychotic medication in AH patients, as there might well be no psychosis.
To improve wellbeing of AH sufferers, PCT suggests cognitive therapies that focus on decreasing both rumination and activation of negative schemas. Waters and colleagues (2017) suggest psychosocial interventions should be the treatment of first choice, including CBT, trauma-related therapy, and psychoeducation. Additionally, mindfulness therapy has been shown to improve mood and enhance connectivity within brain regions involved in AH (Nazimek, Hunter, & Woodruff, 2012).
Hugdahl and Sommer (2018) propose a “Levels of Explanation” understanding of psychotic AH, integrating perspectives from the cultural through to the molecular and genetic levels. An adapted version incorporating non-psychotic AH, presented below, could usefully inform future AH discussions beyond the dominant neuroscience paradigm, and the “hearing voices/psychosis” stigma.
Auditory Hallucinations are distressing events commonly treated with antipsychotic drugs. Yet AH is often experienced by non-psychotic populations, and treatment response is often poor. Newer medications such as agomelatine may provide better results by improving neuroplasticity and white matter connectivity. PCT suggests AH result from states of stress and/or anxiety which permit overly influential prior expectations to suppress normal functioning of the prediction-error signal, allowing internally generated percepts to seem real. This potentially explains how AH arises in stress-related disorders and otherwise healthy populations. Accordingly, psychosocial interventions such as CBT and mindfulness should be among first-choice treatments. The Levels of Explanation framework proposed by Hudgdahl and Sommer (2018) is a useful tool for advancing and integrating understanding and treatment of AH.
References
Akiki, T. J., Averill, C. L., Wrocklage, K. M., Schweinsburg, B., Scott, J. C., Martini, B., . . . Abdallah, a. C. (2017, Aug). The Association of PTSD Symptom Severity With Localized Hippocampus and Amygdala Abnormalities. 1(10). doi:https://doi.org/10.1177/2470547017724069
Altman, M. (2018). Introduction to Psychology. Retrieved from https://app.tophat.com/e/650871
APA. (2013). American Psychiatric Association: Diagnostic and Statistical Manual of Mental Disorders. American Psychiatric Publishing. Retrieved from http://ebookcentral.proquest.com/lib/cdu/detail.action?docID=1811753
Badcock, J. C., & Hugdahl, K. (2012). Cognitive mechanisms of auditory verbal hallucinations in psychotic and non-psychotic groups. Neuroscience and Biobehavioral Reviews, 36, 431-438. doi:10.1016/j.neubiorev.2011.07.010
Badcock, J., Paulik, G., & Maybery, M. (2011). The role of emotion regulation in auditory hallucinations. Psychiatry Research, 185, 303-308. doi:10.1016/j.psychres.2010.07.011
Baldaçara, L., Zugman, A., Araújo, C., Cogo-Moreira, H., Lacerda, A. L., Schoedl, A., . . . Jackowski, A. P. (2014). Reduction of anterior cingulate in adults with urban violence- related PTSD. Journal of Affective Disorders, 168, 13-20. doi:https://doi.org/10.1016/j.jad.2014.06.036
Bobes, J., Gilbert, J., Ciudad, A., Alvarez, E., Canas, F., Carrasco, J., & al, e. (2003). Safety and effectiveness of olanzapine versus conventional antipsychotics in the acture treatment of first-episode schizophrenic inpatients. Progress in Neuropsychopharmacology & Biological Psychiatry, 27(3), 473-481. doi:10.1016/S0278-5846(03)00035-6
Bohlken, M., Hugdahl, K., & Sommer, I. (2017). Auditory verbal hallucinations: neuroimaging and treatment. Psychological Medicine, 47, 199-208. doi:10.1017/S003329171600115X
Clark, A. (2013, Jun). Whatever next? Predictive brains, siguated agents, and the future of cognitive science. Behavioural and Brain Sciences, 36(3), 181-204. doi:10.1017/S0140525X12000477
De Berardis, D., Conti, C., Marini, S., Ferri, F., Iasevoli, F., Valchera, A. .., . . . Di Giannantonio, M. (2013). Is There a Role for Agomelatine in the Treatment of Anxiety Disorders? A Review of Published Data. International Journal of Immunopathology and Pharmacology, 26(2), 299-304. doi:10.1177/039463201302600203
Dudle, R., Dodson, G., Sarll, G., Halhead, R., Bolas, H., & McCarthy-Jones, S. (2014). The effect of arousal on auditory threat detection and the relationship to auditory hallucinations. Journal of Behaviour Therapy and Experimental Psychiatry, 45, 311-318. doi:dx.doi.org/10.1016/j.jbtep.2014.02.002
Fitzgerald, J. M., Phan, K. L., Kennedy, A. E., Shankman, S. A., Langenecker, S. A., & Klumpp, H. (2017, Aug). Prefrontal and amygdala engagement during emotional reactivity and regulation in generalized anxiety disorder. Journal of Affective Disorders, 218, 398-406. doi:10.1016/j.jad.2017.05.013
Friston, K. (2005). Hallucinations and perceptual inference. Behavioural Brain Science, 28, 764-766.
Goldstein, E. B. (2018). Cognitive Psychology: Connecting Mind, Researach and Everyday Experience (4th ed.). Ch. 3, Perception: Cengage Learning.
Hoskin, R., Hunter, M., & Woodruff, P. (2014). The effect of psychological stress and expectation on auditory perception” A signal detection analysis. The British Journal of Psychology, 105, 524-546. doi:10.1111/bjop.12048
Hugdahl, K. (2017). Auditory Hallucinations as Translational Psychiatry: Evidence from Magnetic Resonance Imaging. The Balkan Medical Journal, 34, 504-513. doi:10.4274/balkanmedj.2017.1226
Hugdahl, K., & Sommer, I. (2018). Auditory Verhal Hallucinations in Schizophrenia from a Levels of Explanation Perspective. Schizophrenia Bulletin, 44(2), 234-241. doi:10.1093/schbul/sbx142
Hugdahl, K., & Sommer, I. E. (2018). Auditory Verbal Hallucinations in Schizophrenia from a Levels of Explanation Perspective. Schizophrenia, 44(2), 234-241. doi:10.1093/schbul/sbx142
Hugdahl, K., Craven, A., Nygård, M., Løberg, E., Berle, J., Johnsen, E., . . . Ersland, L. (2015). Glutamate as a mediating transmitter for auditory hallucinations in schizophrenia: a H MRS study. Schizophrenia Research, 161, 252-260.
Jeon, S. W., & Kim, Y.-K. (2016). Molecular Neurobiology and Promising New Treatment in Depression. International Journal of Molecular Sciences, 17(3). doi:10.3390/ijms17030381
Kring, A. (2018). Abnormal Psychology (First Australasian ed.). Hoboken, United States: John Wiley & Sons.
Logue, M. W., van Rooij, S. J., Dennis, E. L., Davis, S. L., Hayes, J. P., Stevens, J. S., . . . Peverill. (2018, Feb). Smaller Hippocampal Volume in Posttraumatic Stress Disorder: A Multisite ENIGMA-PGC Study: Subcortical Volumetry Results From Posttraumatic Stress Disorder Consortia. Biological Psychiatry, 83(3), 244-253. doi: 10.1016/j.biopsych.2017.09.006
Mertin, P., & O’Brien, N. (2013). High emotional arousal and failures in reality monitoring: Pathways to. Scandinavian Journal of Psychology, 54, 102-106. doi:10.1111/sjop.12019
Nazimek, J., Hunter, M., & Woodruff, P. (2012). Auditory hallucinations: Expectation-perception model. Medical Hypotheses, 78, 802-810. doi:http://dx.doi.org/10.1016/j.mehy.2012.03.014
Nazimek, J., Hunter, M., Hoskin, R., Wilkinson, I., & Woodruff, P. (2013). Neural basis of auditory expectation within temporal cortex. Neuropsycholgia(51), 2245-2250. doi:http://dx.doi.org/10.1016/j.neuropsychologia.2013.07.019
Prochwicz, K., Klosowska, J., & Sznajder, D. (2018). The role of expressive suppression in hallucinatory-lie and delusion-like experiences. Findings in a non-clinical sample. 82, 121-127. doi://doi.org/10.1016/j.comppsych.2018.02.001
Sommer, I. E., Koops, S., & Blom, J. D. (2012, Feb). Comparison of auditory hallucinations across different disorders and syndromes. Neuropsychiatry, 2(1), 57-73. doi:http://dx.doi.org.ezproxy.cdu.edu.au/10.2217/npy.12.2
Sommer, I. E., Slotema, C. W., Daskalakis, Z. J., Derks, E. M., Blom, J. D., & Gaag, M. v. (2012, June). The Treatment of Hallucinations in Schizophrenia Spectrum Disorders. Schizophrenia Bulletin, 38(4), 704-14. doi:doi-org.ezproxy.cdu.edu.au/10.1093/schbul/sbs034
Toh, W. L., Thomas, N., & Rossell, S. L. (2015). Auditory verbal hallucinations in bipolar disorder (BD) and major depressive disorder (MDD): A systematic review. Journal of Affective Disorders, 184, 18-28. doi:dx.doi.org/10.1016/j.jad.2015.05.040
Waters, F., Blom, J., Jardri, R., Hugdahl, K., & Sommer, I. (2018). Auditory hallucinations, not necessarily a hallmark of psychotic disorder. Psychological Medicine, 48, 529-536. doi:10.1017/S0033291717002203
