Content uploaded by Don Condie
Author content
All content in this area was uploaded by Don Condie on Mar 18, 2015
Content may be subject to copyright.
ionizing radiation. To investigate the neuronal substrate in-
volved in the personality switches of DID, we conducted an
fMRI study on a woman with DID while she was switching
personalities.
METHODS
The subject is a 47-year-old, right-handed, Caucasian
woman. She was a victim of severe childhood physical and
sexual abuse and was diagnosed with DID and PTSD ac-
cording to DSM-IV criteria.
1
The history and diagnosis had
been recognized since her early adulthood, and she had re-
ceived long-term psychotherapy with one of the authors.
A complete description of the study was provided to the
subject, and written informed consent was obtained. We
identified the alter personality that appeared most often
during psychotherapy; it consistently appeared upon re-
quest. Simulations of personality switch were performed to
prepare the patient for the fMRI study. To identify the
switching period, the subject was instructed to press a spe-
cific button when she started to switch and again when she
fully became the alter or returned to the native personality.
The personality switch was confirmed by the observation of
the subject’s long-term psychiatrist during both the simula-
tion and the fMRI study.
A volumetric study was performed according to the
method of Jack and colleagues;
5,6
the fMRI study used scan-
ning methods and data analysis techniques similar to those
reported by Stern and colleagues
7
at Massachusetts General
Hospital. The subject was scanned using a 1.5-T Signa de-
vice (General Electric Co., Milwaukee, Wisconsin; modified
by Advanced NMR Systems, Wilmington, Massachusetts). A
pillow and foam padding were used to minimize head move-
ment. A set of sagittal localizer scans was performed first to
orient 13 contiguous axial oblique slices from vortex to cere-
bellum. Functional scanning used an asymmetrical spin-
echo imaging sequence, which is sensitive to signal-intensity
changes arising from small alterations in blood-oxygen level.
A gradient-echo T2-weighted sequence was performed using
NEUROSCIENCE
Editor: Joseph T. Coyle, MD
Functional Magnetic Resonance Imaging of
Personality Switches in a Woman with
Dissociative Identity Disorder
Guochuan E. Tsai, MD, PhD, Donald Condie, MD, Ming-Ting Wu, MD, and I-Wen Chang, BA
Dissociative identity disorder (DID) is a controversial condi-
tion characterized by fluctuating constellations of affective,
memory, and behavioral symptoms that are associated with
a changing sense of personal identity.
1
Clinically, DID is of-
ten associated with posttraumatic stress disorder (PTSD).
Psychodynamically, it is considered rooted in childhood trau-
matic experiences and is usually an extreme form of defense
against overwhelming abuse.
Researchers have recently reported hippocampal volume
to be decreased in persons with PTSD secondary to child-
hood abuse.
2
The hippocampus is sensitive to stress and ov-
erarousal,
2
which individuals with DID—like those with
PTSD—often encounter in early life. We performed a volu-
metric study of the hippocampus in a patient with comorbid
DID and PTSD for comparison with the literature on PTSD.
Very little is known about the brain substrate involved in
DID, particularly the mechanism of personality switch. Neu-
rophysiological study of DID is limited due to the dynamic
nature of the disorder. Although measurable neurophysio-
logical differences among personalities in DID have been re-
ported, the results have been inconsistent and often mixed.
3
Unlike other brain mapping techniques, functional mag-
netic resonance imaging (fMRI) has high temporal and spa-
tial resolution,
4
so it is an ideal tool to study neuropsychiat-
ric disorders with dynamic symptoms, such as DID. And
with fMRI—in contrast to procedures such as positron emis-
sion tomography and single photon emission tomography—
a subject can be studied repeatedly without the concern of
119
Supported in part by a NARSAD Young Investigator Award and a
Stanley Foundation Research Award (Dr. Tsai).
Reprint requests: Guochuan Tsai, MD, PhD, Laboratory of Molecu-
lar and Psychiatric Neuroscience, Mailman Research Center,
McLean Hospital, 115 Mill St., Belmont, MA 02478.
q 1999 President and Fellows of Harvard College
Harvard Review of Psychiatry 1999;7:119–122.
the BOLD (blood oxygen level–dependent) technique (TE 5
70 ms, TR 5 2 ms, flip angle 5 90°, acquisition matrix 5 64
3 128).
The subject wore headphones, and the instructions for
switching personalities or guided imagination were played
to her on a tape. As in the simulation session, she heard the
instruction of personality switch, “switch to . . . now.” As soon
as she pressed the button to indicate that she had fully
switched to the alter (native) personality, she was instructed
to switch back to the native (alter) personality. FMRI studies
were performed while she was switching back and forth be-
tween native and alter personalities. To control for volitional
changes, a guided imagination of an irrelevant personality
using a closely matched but emotionally neutral narrative
was performed using recorded instruction, as described by
Cahill and colleagues
8
(see Table 1). To determine test-retest
reliability, the subject was rescanned 4 weeks after the ini-
tial study. The psychiatrist participated in the fMRI scan-
ning procedures to ensure the clinical safety of the subject.
After motion correction, the personality switch–induced
changes in fMRI signal intensity were assessed using a
voxel-by-voxel nonparametric analysis procedure to com-
pare the images acquired during personality switching to
those of the originating personality state.
7
Kolmogorov-
Smirnov statistics were computed for each voxel to assess
the point(s) of maximal difference between the two esti-
mated distribution functions during the control (either na-
tive or alter personality) state and the experimental (switch)
phase (see Figure 1).
RESULTS
The total intracranial volume of the subject (1397.71 cm
3
)
was within normal range. The hippocampal volume (right,
1.09 cm
3
; left, 1.15 cm
3
), however, was significantly smaller
than reported values for normal female adults
6
obtained by
the same technique (right, 2.8 6 0.1 cm
3
; left, 2.5 6 0.1 cm
3
);
it was similar to the volume found in persons with Alzhei-
mer’s disease, who have significant loss of hippocampal
tissue.
5
In the initial investigation we performed three personal-
ity-switch studies, with a total of 12 cycles of switching from
the native to the alter personality and back again. The
switch phase from native personality to alter averaged 27.6
6 3.0 seconds (mean6standard error); from alter back to na-
tive personality, 32.3 6 2.2 seconds.
FMRI demonstrated changes of brain activity during per-
sonality switches (see Figure 1). The switch from native to
alter personality involved bilateral hippocampal inhibition
(less activity during switching than in the native state), with
inhibition stronger on the right side (Figure 1A,B). The right
parahippocampal and medial temporal regions were also in-
hibited (Figure 1A), as were small regions of the substantia
nigra (Figure 1A) and globus pallidus (Figure 1B). In con-
trast, the switch back toward the native personality involved
only right hippocampal activation (i.e., more activity during
switching than in the native state) (Figure 1C).
No significant activation or inhibition was observed in
other brain regions during either direction of personality
switch. The guided imagination of an irrelevant personality
did not evoke a significant change of fMRI signal. The re-
scanning results were similar to those of the initial study.
DISCUSSION
We found bilateral reduction of hippocampal volume in this
subject with comorbid PTSD and DID, consistent with the
reported reduction of hippocampal volume in patients with
PTSD secondary to childhood trauma.
9
In addition, fMRI
during volitionally induced personality switch showed
changes in hippocampal and medial temporal activity corre-
lated with the switch, suggesting that personality switch
may result from changes in hippocampal and temporal func-
tion. These findings are consistent with the few existing neu-
rophysiological studies of DID,
10–12
in which single photon
emission computed tomography and electroencephalogra-
phy have revealed changes in temporal regions between dif-
ferent personalities in persons with this disorder.
Animal studies and human clinical evidence suggest that
stress has long-term effects on the brain regions involved in
memory, particularly the hippocampus.
2,9
In this subject, re-
duced hippocampal volume may be relevant to the attenua-
tion of hippocampal activity during personality switch to-
ward the alter personality.
In contrast to the inhibition seen during the switch to-
ward the alter personality, activation was observed in the
Harvard Rev Psychiatry
July/August 1999
120 Tsai et al
TABLE 1. Instructions for Personality Switch and Guided Imagination
Alter personality Irrelevant personality
Switch to “Guardian” now. Don’t switch.
Imagine that you are an 8-year-old girl named Guardian. Imagine that you are an 8-year-old girl named Player.
You like to watch what happens with [patient’s name]. You like to play with a neighbor girl named Helen.
[Patient’s name] was sexually abused by her father and has been Helen gets along with her father, as in an ordinary family.
very sad throughout her life.
Memory Scale. Scores on all of the subscales were in the av-
erage range. This finding differentiated our subject from pa-
tients with memory disorders. It was also consistent with
the notion that the correlation between deficits in short-term
verbal memory and decreased right-hippocampal volume in
patients with combat-related PTSD does not extend to pa-
tients with PTSD related to childhood abuse.
9,15
Future
study including these clinical, neuropsychological, and
functional-brain-imaging variables will help to delineate the
relationship between PTSD and DID.
Second, in our study personality switching was initiated
by conscious effort. The capacity for volitional elicitation and
reintegration of an alter personality in our subject is not
characteristic of DID patients in general. This volitional pro-
cess may be distinct from spontaneous pathological switch-
ing and cannot be generalized to DID patients who present
only with spontaneous switching. In fact, screening of DID
patients for the study indicated that only a minority of them
could successfully and reliably signal the personality status
and the beginning of a personality switch. The majority,
early in their recovery to personality consolidation, switched
randomly into different alters and failed reliably to report
the change of personality status. Thus, stage of reintegra-
tion recovery may be a relevant study variable and further
limits the generalizability of our findings.
Third, conscious recollection and episodic retrieval are
associated with changes in hippocampal blood flow.
15,16
In
addition, affective and emotional states can affect hippocam-
pal activity. The emotionally neutral, irrelevant personality
control that we performed did not address the issue of differ-
right hippocampus when the subject was returning from the
alter to the native personality. Such activation suggests that
a hippocampus-mediated process is also instrumental in the
subject’s return to the native personality. This adds to the
intriguing possibility that the hippocampus may be involved
in personality switch and dissociation in DID. Activation of
the hippocampus associated with a return to the native per-
sonality suggests a possible neurophysiological basis for an
empirical treatment principle in DID—i.e., consolidating al-
ter personalities by facilitating processing and integration
of early traumatic memories.
In addition to changes in medial temporal and hippocam-
pal activity, nigrostriatal-system inhibition was also asso-
ciated with personality switch (see Figure 1A,B). The hip-
pocampus and associated medial temporal cortical areas
(including entorhinal, perirhinal, and parahippocampal re-
gions) are important neuronal structures for declarative
memory (conscious memory of specific events).
13
The nigro-
striatal system is essential for the gradual, incremental
learning of association (nondeclarative memory).
14
Taken to-
gether, our findings suggest that structures of both declara-
tive memory and nondeclarative memory may be involved in
the personality switches of DID.
Our study has several limitations. First, our subject had
comorbid PTSD and DID; thus our results are not specific for
DID. Clinically, it is difficult to identify a subpopulation of
patients with DID who do not also have symptoms consis-
tent with a diagnosis of PTSD.
Interestingly, our subject was not memory-impaired, as
indicated by a normal score (106) on the general Wechsler
Harvard Rev Psychiatry
Volume 7, Number 2
Tsai et al 121
FIGURE 1. Changes in brain activity during switch between native and alter personalities. The statistical maps comparing the control
phase (native or alter personality) and the switching phase were transformed to 2log(p) maps and translated into a gray scale before being
superimposed over the high-resolution T1-weighted images for anatomic localization. Areas with statistically significant fMRI signal-
intensity changes (p , 0.001) are indicated by arrows. A. Inhibition of right hippocampus, parahippocampus, medial temporal structures,
and substantia nigra during switch from the native to the alter personality (control: native phase). B. Inhibition of the hippocampus
(bilateral, but right side stronger than left) and globus pallidus (right) during switch from the native to the alter personality (control:
native phase). C. Activation of a small region of the right hippocampus during switch from the alter to the native personality (control:
alter phase). R, right; L, left; g, globus pallidus; h, hippocampus; p, parahippocampal gyrus; s, substantia nigra.
ent affective states induced consciously. That is, affectively
charged volitional recall could be contributing to our find-
ings. Different affective states in patients with DID have not
yet been studied with fMRI.
Finally, our findings could reflect a compensatory re-
sponse of the hippocampal regions to a primary functional
deficit that is not detected by fMRI. Another interpretation
of our results is that the observed change in brain activity
during personality switching reflects the difference between
disparate personalities rather than the process of switching
per se. Since we performed the study for only two personali-
ties, we cannot resolve this issue. A large case series study
might elucidate the common mechanism of switching, if
one exists.
REFERENCES
1. American Psychiatric Association. Diagnostic and statistical
manual of mental disorders. 4th ed. Washington, DC: Ameri-
can Psychiatric Association, 1994.
2. Bremner JD, Randall P, Vermetten E, Staib L, Bronen RA,
Mazure C, et al. Magnetic resonance imaging–based measure-
ment of hippocampal volume in posttraumatic stress disorder
related to childhood physical and sexual abuse—a prelimi-
nary report. Biol Psychiatry 1997;41:23–32.
3. North CS, Ryall JEM, Ricci DA, Wetzel RD. Multiple personal-
ities, multiple disorders: psychiatric classification and media
influence. New York: Oxford University Press, 1993.
4. Buonocore MH, Hecht ST. Functional magnetic resonance im-
aging depicts the brain in action. Nat Med 1995;1:379–81.
5. Jack CR Jr, Peterson RC, O’Brien PC, Tangalos EG. MR-based
hippocampal volumetry in the diagnosis of Alzheimer’s dis-
ease. Neurology 1992;42:183–8.
6. Jack CR Jr, Twomey CK, Zinsmeister AR, Sharbrough FW, Pe-
terson RC, Cascino GD. Anterior temporal lobes and hippo-
campal formations: normative volumetric measurements from
MR images in young adults. Radiology 1989;172:549–54.
7. Stern CE, Corkin S, Gonza
´
lez RG, Guimaraes AR, Baker JR,
Jennings PJ, et al. The hippocampal formation participates
in novel picture encoding: evidence from functional magnetic
resonance imaging. Proc Natl Acad Sci USA 1996;93:8660–5.
8. Cahill L, Prins B, Weber M, McGaugh JL. b-Adrenergic acti-
vation and memory for emotional events. Nature 1994;371:
702–4.
9. Sapolsky RM. Why stress is bad for your brain. Science 1996;
273:749–50.
10. Hughes JR, Kuhlman DT, Fichtner CG, Gruenfeld MJ. Brain
mapping in a case of multiple personality. Clin Electroenceph-
alogr 1990;21:200–9.
11. Mesulam MM. Dissociative states with abnormal temporal
lobe EEG: multiple personality and the illusion of possession.
Arch Neurol 1981;38:176–81.
12. Saxe GN, Vasile RG, Hill TC, Bloomingdale K, Van der Kolk
BA. SPECT imaging and multiple personality disorder. J Nerv
Ment Dis 1992;180:662–3.
13. Zola-Morgan S, Squire LR. Neuroanatomy of memory. Annu
Rev Neurosci 1993;16:547–63.
14. Knowlton BJ, Mangels JA, Squire LR. A neostriatal habit
learning system in humans. Science 1996;273:1399–402.
15. Schacter DL, Alpert NM, Savage CR, Rauch SL, Albert MS.
Conscious recollection and the human hippocampal formation:
evidence from positron emission tomography. Proc Natl Acad
Sci USA 1996;93:321–5.
16. Schacter DL, Uecker A, Reiman E, Yun LS, Bandy D, Chen K,
et al. Effects of size and orientation change on hippocampal
activation during episodic recognition: a PET study. Neurore-
port 1997;8:3993–8.
Harvard Rev Psychiatry
July/August 1999
122 Tsai et al
