Saturday, July 14, 2012

Brain Responses to Virtual Reality-Induced Hallucinations in Schizophrenia

What is it like to experience the frightening auditory and visual hallucinations characteristic of schizophrenia? Yellowlees and Cook (2006) developed a virtual reality program in Second Life based on interviews with schizophrenic patients. The researchers used this as a tool to educate the general public about schizophrenia, in order to increase understanding and reduce stigma. A video sample of the program can be viewed below.

As you can see, these hallucinations are straight out of a horror movie or a terrible nightmare, except they reflect the reality of living with schizophrenia:
  • Multiple voices, occasionally overlapping, criticizing the user
  • A newspaper in which the word “death” would stand out in a headline
  • A floor that would fall away, leaving the user walking on stepping stones above a bank of clouds
  • A television that would play a political speech, but then criticize the user and encourage suicide
  • A gun that would appear under a cone of light and pulse, with associated voices telling the user to take the gun and commit suicide
  • A mirror in which a person's reflection would appear to die, becoming gaunt with bleeding eyes
The authors also provide information about accessing the Virtual Hallucination environment directly.

Persons with other psychiatric disorders may be plagued by voices saying they're worthless and directing them to commit suicide, but the voice is a self-deprecating internal monologue and clearly identified as their own (as in nonpsychotic unipolar and bipolar depression). The issue in schizophrenia is one of reality monitoring, so that internal thoughts and impulses are interpreted as external to the self.

The most common type of hallucination is hearing voices. To determine which brain regions are implicated, a number of neuroimaging studies have scanned participants with schizophrenia while they are actively experiencing auditory hallucinations, compared to the non-hallucinating state (Allen et al., 2008; Kompus et al., 2011; Jardri et al., 2011). A common finding is increased activation of auditory cortex in the absence of external stimulation, along with greater activity in Broca's area (speech production) and the medial temporal lobe (memory). One interpretation of this pattern is that memory retrieval triggers aberrant auditory perceptual experiences. Another is that inner speech is attributed to external sources due to defective self-monitoring.

A new study by Kim and colleagues (2012) took a different approach. They constructed a virtual reality environment in the scanner to produce the illusion of burning flames, and compared the neural responses of schizophrenic and control participants. The experimental setup is shown below.

Fig. 1A (adapted from Kim et al., 2012). The virtual flame illusion. A participant could see his/her body through a head-mounted display (HMD) during the “flame off” block and watched a superimposed, animated image of a virtual flame on the right or left index finger during the “flame on” block.

The participants were 16 schizophrenic patients with mild to moderately severe symptoms and 17 controls. They were instructed that the purpose of the experiment was to examine the brain's response to “observing the body.” However, they were not informed about the potentially frightening illusion:
Participants were not told about the virtual flame and were instructed to observe their body without closing their eyes. As shown in Fig. 1, the experiment used a blocked paradigm and consisted of two conditions: 1) a ‘flame off’ block (30 s), during which only a real-time body image was presented, and 2) a ‘flame on’ block (16 s), during which the virtual flame was generated by a computer in real-time and superimposed on the participant's index finger. The blocks were alternatively repeated 8 times, and the presentation of the flame on the left or the right finger was counterbalanced.

Fig. 1B (adapted from Kim et al., 2012). fMRI task sequence. The experiment used a blocked design and alternated between ‘flame off’ blocks (30 s) and ‘flame on’ blocks (16 s).

Lest you think this situation skirts the boundaries of unethical (as I did), the study was approved by the local institutional review board and subjects signed [semi-]informed consent statements.

After the fMRI session was over, the participants filled out a questionnaire which indicated (A) the strength of their reactions from 1 (“not at all”) to 7 (“extremely strong”), and (B) how much their feelings changed when the blocks were repeated, rated from 1 (“severely attenuated”) to 7 (“severely augmented”):
After scanning, most participants reported that they initially felt the ‘flame on their finger,’ but then the feeling disappeared after realizing that the flame was not real. . . . ...both patient and control groups showed similar subjective responses to the task stimuli: moderate strength in feeling the flame (4.6 ± 1.9 and 3.8 ± 1.8, respectively) and slight attenuation in flame strength over time (3.3 ± 1.8 and 2.6 ± 1.3, respectively). [The group differences were not statistically significant.]

The data analysis strategy went beyond the standard boxcar comparison between "flame on" and "flame off." Instead, the authors...
...considered that the process of virtual flame-specific learning (i.e., gaining insight into the reality of a visual image) might be reflected as a linear or quadratic function of fMRI signal changes. A linear function could reflect repetition enhancement, sensitization/repetition attenuation, or habituation to the stimulus. A quadratic function could reflect transitions between repetition enhancement and repetition attenuation.

The brain activation differences between groups were not all that spectacular, once you discard all the p<.001 uncorrected regions that were reported in Table 2. What was left?
...only five areas including the left anterior prefrontal cortex, left occipito-temporal junction, left occipital gyrus, right amygdala, and left cerebellum were included. As depicted in Fig. 2, the control group demonstrated transitions from repetition enhancement to attenuation in these five brain regions, in contrast to the lack of enhancement and attenuation in the patient group [i.e., a flat response].

Fig. 2 (Kim et al., 2012). Changing patterns of brain activity related to the virtual flame across time in patients with schizophrenia and in healthy controls. The selected regions were significant in the two-sample tests with the quadratic repetition-variant response at p < 0.05, FDR-corrected: the anterior prefrontal cortex (APFC), occipito-temporal junction (OTJ), occipital cortex (OC), amygdala (AMG), and cerebellum (CER). Rt., right; and Lt., left

Have we come away with a better understanding of the neural processes involved in gaining insight into unreality? Becoming aware of the fact that an unexpected and alarming visual illusion isn't real differs from internally generated hallucinations, of course, but the authors suggest that:
Patients with schizophrenia may use a salience-related region1 instead of reality monitoring-related regions [anterior medial PFC] to react to the unusual stimuli, and this peculiarity of the neural processes may be related to vulnerability to psychosis.


1 The salience-related region (anterior cingulate cortex) did not survive FDR correction for multiple comparisons.


Allen P, Larøi F, McGuire PK, Aleman A. (2008). The hallucinating brain: a review of structural and functional neuroimaging studies of hallucinations. Neurosci Biobehav Rev. 32:175-91.

Kim JJ, Ku J, Lee H, Choi SH, & Kim IY (2012). Distinct neural responses used to gain insight into hallucinatory perception in patients with schizophrenia. Journal of psychiatric research PMID: 22770670

Kompus K, Westerhausen R, Hugdahl K. (2011). The "paradoxical" engagement of the primary auditory cortex in patients with auditory verbal hallucinations: a meta-analysis of functional neuroimaging studies. Neuropsychologia 49:3361-9.

Jardri R, Pouchet A, Pins D, Thomas P. (2011). Cortical activations during auditory verbal hallucinations in schizophrenia: a coordinate-based meta-analysis. Am J Psychiatry 168:73-81.

Yellowlees PM, & Cook JN (2006). Education about hallucinations using an internet virtual reality system: a qualitative survey. Academic Psychiatry, 30 (6), 534-9. PMID: 17139026

When we sit at a table there's fire between the guests
When your hands don't touch there's sand in your face  
And fire under your nails  
Nobody knew so nobody cared
 Nobody knew so nobody cared
Nobody knows

------Throwing Muses

Subscribe to Post Comments [Atom]


At July 15, 2012 3:57 AM, Blogger Neuroskeptic said...

The Second Life thing is pretty disturbing, mainly because of the voices though. The graphics are just too crude to be believable.

At July 15, 2012 10:18 PM, Blogger Evening Pilgrim said...

I used to work for a company that developed digital audio encoding and decoding equipment. The goal was always to find inventive ways of generating rich audio via minimal signal input. In the early days if surround sound something they discovered was that the ability to recognize the direction from which a sound is coming is a learned skill. Sounds sound different depending upon how they bounce off the structures of the ear before reaching the ear drums. So a sound coming from behind a person actually sounds somewhat different as a consequence of it bouncing toward the ear drums, than sound coming from other directions. And humans learn to recognize those differences at an early age.

The audio company I worked for studied the qualities of sounds that reached the ear drums from various directions and developed ways of reproducing those qualities. Eventually they were able to mixed sounds that mimicked “from behind” characteristic with sounds as we are accustomed to hear them coming from in front of us. The result was “virtual surround sound” which came out of two speakers in front of the listener but sounded like it included sounds coming from other directions.

Let’s imagine that physical vibrations have been translated to signals within the brain, and interpretation has begun. At that point habits of expectation tell us that people across the room in front of us are talking quietly and someone behind us is shouting. But how do we know those sounds were ever physical vibrations? What if they were always signals…perhaps a sort of recording…interjected into the audio interpreting areas of the brain by… “By” may or may not be an appropriate word, since it suggests a willful entity taking action. “Via” is simpler, and might work equally well, or as an alternative. Signals interjected into the audio interpreting areas of the brain via a loop that includes areas of the brain that manage speech. If established early in life such a loop might acquire sufficient structural integrity to permit significant signal strength. And from the beginning those signals would probably sound different from the sound of one’s own voice.

At July 16, 2012 1:46 AM, Anonymous Anonymous said...

What is the difference between a hallucination and a dream?

At July 29, 2012 3:38 PM, Blogger Raph said...

This comment has been removed by the author.

At July 29, 2012 3:43 PM, Blogger Raph said...

I actually work in Psychosis research, and even I found that video extremely disturbing!
The crude graphics add to it in a way, like how the shoddy stop-motion graphics in Jason and the Argonauts are far scarier than anything computer generated today.

Could I be shameless and send you a post from my blog about consciousness? It's unrelated I know, but I wanted to send something academic-ish that I had written to someone I respect (I hear that's a done thing).



Post a Comment

Links to this post:

Create a Link

<< Home

eXTReMe Tracker