THALAMIC THOUGHT DISORDER:
ON BEING “A BIT ADDLED”*
Anjan Chatterjee1, Robert Yapundich1, Mark Mennemeier2, James M. Mountz3,
Chakri Inampudi3, Jullie W. Pan1,4, Galen W. Mitchell1
(Departments of 1Neurology, 2Rehabilitation Medicine, 3Radiology and 4the Center for
Nuclear Imaging Research, The University of Alabama in Birmingham, Birmingham,
Alabama, USA)

ABSTRACT
Humans can generate and maintain relatively coherent trains of thought in natural
discourse. The neural mediation of this ability and the phenomenology of its breakdown are
not well understood. We report a case of a woman with paramedian thalamic strokes involving
the mammillothalamic tract, intralaminar nuclei, parts of the dorsomedial and ventral lateral
nuclei bilaterally. She presented with a dense amnesia and confusion typical of the syndrome
of bilateral paramedian thalamic infarcts. Her Tc-99m HMPAO brain SPECT scan showed
decreased thalamic and basal ganglia blood flow. General diminution of cerebral blood flow
and areas of further diminution in the right frontal, left temporal and left temporoparietal
regions were also observed. Although her amnesia was characteristic of diencephalic amnesia,
her most striking clinical feature was a bizarre, disconnected and at times incoherent speech
output. Analysis of her speech revealed relatively preserved lexical and morpho-syntactic
linguistic production. By contrast, analysis of the macrostructure of her discourse revealed
frequent unpredictable topic shifts that were completely unconstrained by contextual factors.
Many of her shifts were intrusions from previous topics. We interpret her severely disordered
speech output as representing the surface manifestations of a thought disorder (rather than
as a language disorder per se) characterized by an inability to maintain and appropriately
shift themes that normally guide discourse. Median and intralaminar thalamic nuclei appear
to be critical for the neurophysiologic regulation of thalamocortical and striatocortical circuits,
which in turn may be critical for the functional regulation of contextually appropriate
transitions of thought.

INTRODUCTION
Humans can generate and maintain relatively coherent trains of thought.
Themes can be traced across sentences in conversation as ideas are developed
and expressed. Humans can also shift from one set of ideas to others. These
shifts may vary from gradual transitions to abrupt changes; the nature of these
shifts being dictated by pragmatic and contextual settings in which ideas are
conveyed. William James (1890) referred to these stable and dynamic aspects
of thought as “substantive” and “transitive” states of consciousness, respectively.
The current interest in the neuropsychology of consciousness has focused

* Portions of this paper were presented as a poster at the 24th Annual Meeting of the International Neuropsychological
Society at Chicago in February, 1996.

Cortex, (1997) 33, 419-440

420

Anjan Chatterjee and Others

primarily on what James might have considered substantive aspects of
consciousness. Amnesia, blindsight, unilateral neglect, prosopagnosia, and
anosognosia provide important data for our investigations into the neurobiology
of consciousness (Milner and Rugg, 1992). By contrast, little is known of
transitive aspects of consciousness. How are transitions of thought neurally
regulated? How would the breakdown of this regulation be manifested? How
are such questions best pursued empirically? If transitive aspects of thought are
distinct from substantive ones, then these functions might be mediated and
controlled by different neural systems. Brain damage, under some circumstances,
might then impair thought transitions selectively. James (1890) suggested that
dynamic aspects of thought are revealed in the “rhythm of language”. Thus,
an individual’s discourse in relatively unstructured settings might offer insight
into the structure of thought transitions. Impairment of the ability to regulate
thought transitions would be evident in speech, but would not be explained by
disruptions of local features of language (phonology, syntax, lexical-semantics).
Rather, the disordered speech would be the surface consequence of a dynamic
disorder not captured adequately by our understanding of traditional aphasic
disorders.
With these considerations in mind, we report observations in a patient with
bilateral paramedian thalamic infarcts. Patients with bi-thalamic lesions are
amnestic, frequently appear confused, and may produce bizarre confabulations.
This amnesia has been well studied (Bogousslavsky, Regli and Uske, 1988b;
Castaigne, Lhermitte, Buge et al., 1981; Crosson, 1992; Graff-Radford, Tranel,
Van Hoesen et al., 1990; Von Cramon, Hebel and Schuri, 1985). By contrast,
the confusional state in this syndrome has been examined only rarely (Luria,
1977). In this case study we focus on the nature of our patient’s confusion. We
examine the structure of her discourse and consider its implications for the
neuropsychology and neurobiology of transitions in thought.
CASE HISTORY
LB, a 70 year old right-handed woman with a history of hypertension and coronary
artery disease, was found lying on the floor of her home in a confusional state. Despite
being able to answer questions, she could not offer a cogent explanation for why she was
lying on the floor. In the first few days after admission to the hospital, she was somnolent
but arousable. Her speech was dysarthric initially. She was severely amnestic. Her level of
arousal improved after a few days. Despite this improvement, LB remained disoriented to
place and time, and her responses to questions were often bizarre. For example, when asked
where she was she replied South America. When asked how old she was she replied that
she was 17. A few moments later she said she was in her 30s. Her digit span was five digits
forward. She was unable to repeat two digits backwards. When asked to count backwards
from 20, she did so till 11 before talking about something else. She could follow simple
one and two step commands and repeat phrases without difficulty. Her spontaneous speech
was fluent. On verbal fluency she named 11 words beginning with the letter F in 60 seconds.
Some of these were low frequency words such as fearsome, folly and forlorn. She was
unable to generate words for the semantic categories of animals, tools or fruits. This poor
performance seemed to be a result of her quickly switching topics rather than an inability
to generate words. She performed gestures to command of transitive movements (such as
hammer) well. When asked to follow simple simultaneous motor commands, such as “open
your mouth and raise your hand” she performed the movements accurately but would

Thalamic thought disorder

421

perseverate, repeating the movements five or six times till she was asked to stop. By contrast,
when asked to repeat motor sequences such as the Luria gestural sequence of “fist-edgepalm”, she was able to follow and mimic these sequences, but on her own could not produce
the sequence more than once. She could repeat 10 bits of information in a short story, but
was unable to recall the story after five minutes. She could recite the days of the week, the
letters of the alphabet and a short jingle (row, row, row your boat...). Her conversational
discourse was disconnected, confabulatory and at time bizarre (examples in the Appendix).
One week after her admission her visiting daughter woke her and LB immediately began
drilling her on math problems. The daughter (in her 40s) later explained that her mother
used to test her similarly over 35 years before.
LB’s visual acuity was poor because of a prior history of macular degeneration. Her
cranial nerve exam was notable for restricted up-gaze. She did not have motor or
somatosensory deficits. Her clinical evaluation included normal electrolytes, thyroid function
tests, vitamin B12 level, negative syphilis serologies, ANA titres and rheumatoid factor.
Cerebrospinal fluid analysis was normal. The electroencephalogram revealed mild background
slowing. Her hospital course was notable for gradual improvement in her level of arousal,
without improvement in her amnesia or confusion. At no time during her hospital stay was
LB aggressive or hostile towards the hospital staff. She, however, did demonstrate a kind
of inertia and did not herself initiate activities. She required considerable supervision and
constant encouragement to persist in physical therapy tasks. She continued to be severely
amnestic and confused for at least a year after her initial presentation.

Neuroimaging
Magnetic Resonance Imaging (MRI)
A clinical 1.5 Tesla MRI scans revealed bilateral medial thalamic infarcts
in the paramedian artery distribution. To better delineate her diencephalic lesions,
a 4.1 Tesla 256 × 256 T1 weighted MRI scan was obtained using an inversion
recovery gradient echo sequence (Figure 1). The two axial MRI images in which
lesions were evident were matched to relevant images in Haines’ neuroanatomic
atlas (Haines, 1995). This atlas allows correlation of axial brain imaging slices
with neuropathologic sections. The figures relevant to our case, Figures 6-1 to
6-4 from Haines (1995) show thalamic nuclei in relation to basal ganglia
structures at these levels. Based on comparison with these sections, LB’s lesions
involved the mammillothalamic tract, the internal medullary lamina and parts of
the dorsomedial nucleus and ventral lateral nuclei bilaterally. Small portions of
the ventral anterior nuclei may also have been damaged.
Single Photon Emission Computed Topographic (SPECT)
LB was scanned twice, one week and two months after her stroke using the
same protocol (Mountz, Zhang, Liu et al., 1994b). Regional cerebral blood flow
(rCBF) SPECT brain scans were done after intravenous injection of 25 mCi of
Technetium-99m Hexamethylpropyleneamine oxime (Tc-99 HMPAO). The
patient was injected in a quiet environment with eyes closed in a dimly lit room.
After five minutes of tracer uptake the patient was scanned on the ADAC dualhead Genesys Anger Gamma Camera (ADAC laboratories, MILPITAL, CA).
Scanning parameters were 128 stops (64 stops per head), 30 seconds per stop
using a 25 cm field of view resulting in a pixel dimension size of 0.196 cm
on edge. Images were reconstructed using a Butterfield filter back projection

422

Anjan Chatterjee and Others

Fig. 1 – 4.1 Tesla 256 × 256 T1 weighted MRI scan (TR/TE = 2500/17.5 msec) obtained one week
after her stroke. The bilateral paramedian infarcts are evident in both panels.

algorithm (frequency cut-off = 0.22 Nyquest, order 6) (Liu, Harris, Inampudi et
al., 1995). Attenuation correction was performed using the Chang (1978)
algorithm. A reference system was employed during the scan such that the brain
could be sectioned parallel to and sequentially above the canthomeatal line
(Mountz et al., 1994b). Analysis of the basal ganglia and thalami were performed
by co-registration of the brain SPECT scan with the MRI scan using a reference
system technique developed at this institution (Mountz, Wilson, Wolff et al.,
1994a). Counts were characterized as the average counts in the regions of interest
(ROI) normalized to cerebellar and whole brain counts (Mountz, 1991). Cortical
ROIs were generated using a circumferential ROI technique standard at this
institution (Mountz, 1994). Briefly, an outer boundary is computer drawn at the
50% counts per pixel threshold, and an inner boundary is drawn at a distance
eight pixels radially inward. The annulus is subdivided into twelve equal angular
sections for cortical count semi-quantification. The ratios of maximum and
average cortical counts and cerebellar whole brain counts were calculated
identically for both scans.
The first scan showed increased rCBF in the thalamus (left thalamus = 1.05,
right thalamus = 1.1; normal thalamic uptake = 0.96 + 0.04) interpreted as
representing luxury perfusion. The follow-up scan (Figure 2) showed diminished
bi-thalamic perfusion (left thalamus = 0.70, right thalamus = 0.74) and diminished
basal ganglia flow (left = 0.76, right = 0.73; normal 0.94 + 0.05). Basal ganglia
hypoperfusion is compatible with the fact that intralaminar nuclei have major
efferents to the striatum (Groenewegen and Berendse, 1994; Jones, 1985). She also

Thalamic thought disorder

423

Fig. 2 – (a) Sequential Tc-99m HMPAO brain SPECT scan sections (approximately 5 mm thick)
obtained one week post-stroke. The scans progress from the ventral (top left) to the dorsal part of the
brain (bottom right). High tracer uptake were noted in the thalamic regions bilaterally, which when
integrated with the abnormalities on the MRI scan were interpreted as the “luxury perfusion” phase of
infarction.

Fig. 2 – (b) Sequential transverse Tc-99m HMPAO brain SPECT sections (approximately 5 mm
thick) from the patient 2 months post-stroke. Marked reduction of regional cerebral blood flow to the
thalamus is indicative of bilaretal thalamic infarction.

424

Anjan Chatterjee and Others

Fig. 2 – (c) Tc-99m HMPAO brain SPECT scan sections oriented parallel to and 5.5 cm above the
canthomeatal line. The changes in blood flow to the thalamus from the one week post-stroke scan (left)
are contrasted with the marked reduction of blood flow typically associated with infarction on the 2
month blood flow scan section (right).

Fig. 2 – (d) Tc-99m HMPAO brain SPECT scan sections from the one week post-stroke scan
positioned parallel to and 5.5 cm (top left) and 7.5 cm (bottom left) above the canthomeatal line. To the
right of the images are shown the semi-quantitative cortical circumferential profile of the cortical to
cerebellar indices as described in the text. The blue line represents the patient, the green line represents
one standard deviation from age matched normal controls (n = 9), and the red line indicates two
standard deviations from the mean of age-matched normal controls. The x axis represents cortical
regions progressing from the twelve to one o’clock position (cortical region #1) to the eleven to twelve
o’clock position (cortical region #12). The y axis has the units of cortical region counts divided by
cerebellar counts. The semi-quantitative curves show reduction of regional cerebral blood flow in the
left temporal region (top curve region 4 at canthomeatal + 5.5 cm). The bottom curve shows reduction
of left posterior parietal cerebral blood flow.

Thalamic thought disorder

425

Fig. 2 – (e) Tc-99m HMPAO brain SPECT scan sections from the two month post-stroke scan
presented in the format described for figure 2d. The semi-quantitative profile for the level canthomeatal
+ 5.5 cm shows further reduction of left temporal blood flow (region 4) and the right frontal blood flow
(region 12). At canthomeatal + 7.5 there is also further reduction of regional blood flow at the left
parietal lobe (region 4).

had global reductions of cerebral blood flow and further diminution (greater than
two standard deviations from normal age matched sample) in the left temporal,
posterior parietal and right anterior prefrontal cortices.
Behavioral and Neuropsychological Testing
Attempts at formal psychometric testing were compromised by LB’s inability
to persist on tasks. On the Dementia Rating Scale (Mattis, 1976) she scored
44/144, with the following sub-scores: attention 14/37, initiation and perseveration
16/37, construction 0/6, conceptualization 14/39, memory 0/25. We have doubts
about the validity of these scores since she frequently seemed not to be trying
to perform the required task. Further testing of general intelligence was not
pursued since these assessments seemed unable to capture the peculiarities of
her deficit.
LB’s performances on the Wechsler Memory Scale-Revised (Russell, 1975)
and the California Verbal Learning Test (CVLT) (Delis, Kramer, Kaplan et al.,
1987) were characteristic of a dense amnesia. She was unable to recall any
information after a 20 minute delay and did not display any learning on the
CVLT. Responses to the logical memory stories were confabulated and there
were more intrusions than correct responses on the CVLT.

426

Anjan Chatterjee and Others

Language
LB followed simple commands and repeated phrases easily. Confrontation
naming could not be tested adequately because of her poor visual acuity. Two
extended interviews conducted by AC were videotaped. The first (lasting 56
minutes and 43 seconds) was done 11 days after her stroke and the second (43
minutes and 25 seconds) three weeks after her stroke. These interviews followed
the general format of a neurological history and mental status exam, however,
the content drifted from the traditional confines of a clinical interview, as
described below. The taped interviews were transcribed independently by two
raters. Point-to-point agreement by each word was 0.77 for the first interview
(when she was dysarthric) and 0.86 for the second interview. Differences were
decided by AC who conducted the interview, after a review of the transcriptions
and the videotapes. These transcripts served as the corpus from which subsequent
analyses were performed.
Analysis of Syntax
In focusing on the nature of LB’s discourse, we wished to examine the
structure of her language production. To assess her syntactic abilities we adapted
the method described by Saffran et al. to analyze narrative production (Saffran,
Berndt and Schwartz, 1989). Subjects are asked to narrate well known fairy
tales, such as the Cinderella story. From 150 words of these narrations,
grammatical properties of language output can be quantified. Attempts to get
LB to generate fairy tale narratives were unsuccessful (see Appendix). However,
to confirm the general impression of preservation of lexical selection and
structure, two portions of LB’s discourse were selected (one from each session)
in which she talked without interruption for over 150 consecutive words. In
addition, the beginning of the first interview was analyzed similarly despite the
disconnected nature of this part of her discourse. Since this corpus was not
derived from the general domain of telling fairy tales we recognize that the
comparison of LB’s production to these normative data has limitations. However,
our focus was to learn if the microstructure of her production more closely
resembled normal production than agrammatic production. If so, one would not
ascribe her disordered output to fragmentation of local features of language. As
shown in Table I, her performance for proportion of closed class words, noun/
pronoun ratio, presence of obligatory determiners, noun/verb ratio and frequency
of verb inflections was close to normal. In addition, the proportion of words in
sentences, the proportion of well formed sentences, indices of sentence structure
elaboration and embeddings also resembled normal performance where she spoke
without prompts for extended periods. In sections of her discourse comprised of
smaller segments, indices of structural sentence elaborations were reduced.
It is of interest that that impoverished structural elaboration may reflect a
conceptual rather than a syntactic disorder (Berndt, 1991).
Analysis of Discourse Macrostructure
Description (also see Appendix). LB did not violate conventions of turn

Thalamic thought disorder

427

taking in conversation and generally maintained prosodic contours of speech
appropriate to the context of what was being discussed. She made appropriate
eye contact and seemed to behaviorally adopt the interactional style one would
expect of conversation as a joint activity. When not required to produce a verbal
response to simple commands (touch your nose), she performed well. She also
performed simple constrained verbal commands, such as repeating digits or
phrases, accurately. Her replies to simple questions in which her verbal response
was unconstrained fit the form of the question but were often bizarre. For
example, when asked where she was, she replied “South America”. With openended questions her discourse became disorganized. At times she seemed to
start appropriately but quickly drifted off to new topics. The reason for the
content of her speech and topic transitions was usually indecipherable from the
context. At times she seemed ‘stuck in set’. At other times previous topics would
intrude upon her discourse. Fragments of themes would surface unpredictably
and occasionally more than one theme would be fused within single statements.
To better characterize these dynamic features of her disordered speech we
analyzed her topic transitions formally.
Topic Shifting. Garcia and Joanette (1994), following Crow (1983), point out
the insensitivity of traditional aphasia assessment tools to adequately describing
disorders of discourse. Rather, they develop an approach predicated on the notion
that an overarching theme or topic guides conversational discourse. Utterances
cohere to each other by their relevance to the topic. To confirm the clinical
impression that LB’s transitions were severely disordered, we adapted their
method of topic shift analysis. Garcia and Joannette (1994) distinguish
interpretative from sequential issues in such analysis. Interpretative issues are
concerned with the intended meaning of an utterance. Sequential issues are
concerned with connections between utterances and the dynamics of topic shifts.
While the bizarreness of LB’s responses (Appendix) raise interesting questions of
interpretation, we will only touch on these tangentially. We focus mainly on
sequential issues in LB’s discourse.
The text from LB’s conversational discourse was coded by utterances. A
noun phrase or sentence was considered an utterance. Some portions of her
TABLE I

Microstructural analysis adapted from Saffran et al. (1989). Their data from controls and
agrammatic aphasics are included to show LB’s general pattern across both interviews LB’s first
column reports analyses sections of from uninterrupted speech. The second column reports analyses
from the beginning of the first conversation, which includes interrupted discourse.

Proportion closed class words
Noun/pronoun ratio
Noun/verb ratio
Frequency of verb inflection
Determiner/noun ratio
Proportion of words in sentence
Proportion sentences well formed
Sentence structural elaboration
Embedding index

Ag

Cntrl

LB

LB

0.33 (0.08)
8.93 (0.27)
2.93 (0.92)
0.34 (0.29)
0.37 (0.24)
0.43 (0.14)
0.33 (0.22)
1.07 (0.22)
0.28 (0.04)

0.56 (0.04)
1.60 (0.85)
1.18 (0.30)
0.98 (0.05)
1.00 (0)
1.00 (0)
0.91 (0.05)
3.76 (6.00)
0.53 (0.33)

0.67
1.39
1.18
0.85
0.90
0.88
0.92
2.88
0.47

0.46
1.43
0.94
0.92
1.00
0.95
0.90
2.20
0.25

428

Anjan Chatterjee and Others

discourse were considered fragments and not included in the analyses. These
fragments were false starts, unconnected prepositional phrases or isolated words
from which a theme could not be deciphered. Fragments were excluded from
the analysis to be conservative in our estimations of topical shifts. The specific
points at which a topic shift were identified was then coded for the characteristics
of these shifts. These topic shifts were categorized for 1) type of shift, 2) reason
for shift and 3) context of shift. Operational definitions for these categories are
given in Table II. The types of shift were either topic initiation (TI), topic
shading (TS), topic drift (TD), renewal of topic (R), unexpected (U) or conflation
(C). Only TI and TS are considered normal shifts. Reasons for the shifts were
coded as the end of a topic (ET), answering previous question (AQ), reacting
to an outside stimulus (OS) or unknown (U). The context of the shift refers to
whether the topic shift related to information or knowledge shared by both
participants. These were categorized as referring to topics previously initiated
externally by the examiner (ET), topics previously initiated internally by the
patient (IT) and unknown (U).
AC and RY coded topic shifts independently. Differences in coding were
resolved by consensus. The results for both interviews are depicted in Table III
and Figure 3. In summary, in the first interview she produced 1854 words,
405 utterances, and 111 topic shifts. In the second interview she produced
2250 words, 403 utterances and 113 shifts. The average length of her
utterances increased slightly from 4.6 words to 5.6 words per utterance, but
TABLE II

Operational definitions for classification of topic shifts, adapted from Garcia and Joannette (1994)
following Crow (1983). Only classifications applicable to this case are included. We introduced
classifications marked by (*) to better capture LB’s specific topic transitions
1. Type of Shift
1.1 Topic Initiation (TI). This shift represents an attempt to introduce a new topic at the beginning
of a conversation, or after a period of silence.
1.2 Topic Shading (TS). The introduction of a new topic by first establishing its connection with a
topic already on the floor.
1.3 Topic Drift (TD).* The tangential shifting of a topic from the topic on the floor.
1.4 Topic Renewal (TR). A shift back to a previous topic after one or more topics have since
intervened.
1.5 Unexpected (U). The shift to a new topic in a way that is unexpected and is difficult to
understand from the context in which it occurred.
1.6 Conflation (C).* The new topic forms a hybrid combining at least two previously mentioned
distinct topics.
2. Reason for Shift
2.1 End of Topic (ET). When the topic shifts because it is clear that the speaker has nothing more
to say about the previous topic.
2.2 Answering an earlier question (AQ).* The shift can be traced to answering or responding to a
previous question or point despite intervening topics.
2.3 Outside stimulus (OS). The shift in topic is triggered by some environmental stimulus.
2.4 Undetermined (U). The reason for the shift could not be determined.
3. Relation to Context
3.1 Externally Initiated Topic (ET).* A shift to a topic previously introduced by the interviewer.
3.2 Internally Initiated Topic (IT).* A shift to a topic previously introduced by the patient.
3.3 Unknown (U). The context within which the shift occurred could not be determined.

Thalamic thought disorder

429

utterances per shift remained unchanged at 3.6 utterances per shift. Topic
shifts only refer to shifts made by LB and not the interviewer. Most of her
shifts were either completely unexpected or inappropriate renewals of previously
mentioned topics. Very few (less than 5%) of her topics shifts were either
the initiation of or shading into new topics, both considered appropriate and
normal types of shifts. In the vast majority of instances (> 90%) the reason for
her shift was not evident from the context of the discourse. Almost half of
these unexpected shifts introduced novel topics with no previously shared referent,
and the rest were intrusions. The intrusions were about evenly divided between
topics the patient had previously introduced, and topics introduced by the
interviewer.
A Control Group
Although it seemed self-evident that LB’s topic shifting was strikingly
abnormal, we compared her discourse to that of three patients with right
hemisphere strokes. Right brain damage is also associated with disorders of
discourse (Bloom, 1994). These patients were video taped for extended
interviews regarding the awareness of their left sided weakness. Detailed
discussion of these patients are published elsewhere (Chatterjee and Mennemeier,
1996). MR had a discrete right cortical stroke confined primarily to motor cortex.
EM had a large right basal ganglia-internal capsule hemorrhagic stroke. HS
had a large right hemispheric stroke involving most of the distribution of the
middle cerebral artery. Their shifting characteristics are outlined in Table III. In
brief, they initiated far fewer shifts than LB. Aggregating their data, 90% (19 of
21) of the shifts they produced were appropriate as compared to 1 or 2% of
LB’s shifts. These appropriate shifts were invariably considered topic shadings,
in which they introduced personal anecdotal information related to the topic
at hand. Of the two inappropriate shifts, one was a topic drift which also
introduced personal anecdotal information, but was not germane to the topic.
The second was unexpected and was the single shift produced by these patients
which was similar to the shifts LB produced commonly. In summary, these
patients’ discourse throws into relief the severe abnormality of LB’s topic shifting
behavior.

TABLE III

General discourse and shifting characteristics of LB and the three right brain damaged control
subjects

LB 1
LB 2
MR
EM
HS

Words

Utterances

Words/
utterance

Shifts

Utterances/
shift

% normal
shifts

1854
2250
1218
803
1275

405
403
159
134
192

4.6
5.6
7.6
6.0
6.6

111
113
7
4
10

3.6
3.6
22.7
33.0
19.2

2
1
86
100
90

430

Anjan Chatterjee and Others

Fig. 3 – Distribution of topic shift characteristics. Black columns represent the first interview, and
gray columns represent the second. a) Type of shift. TI - topic initiation, TS - topic shading, TD - topic
drift, R - renewal of topic, U - unexpected, C - conflation. See Table II for operational definitions. b)
Reasons for shift. U - unknown, AQ - answering previous question, OS - reacting to an outside stimulus.
See Table II for operational definitions. c) Context of shift. U - unknown, IT - topic previously initiated
by the patient, ET - topic previously initiated by examiner. See Table II for operational definitions.

Thalamic thought disorder

431

DISCUSSION
LB presented with an acute amnesia and a confusional state. Her initial
lethargy and dysarthric speech resolved within a few days. However, her dense
amnesia and confusional state persisted for at least a year after her stroke. She had
minimal insight into her deficits. Six months after her stroke when asked
if she had any problems with her thinking, she laughed and replied that she was
“a bit addled”, but she had been that way all her life. Her MRI scans showed
bilateral medial thalamic lesions in the vascular territories of the paramedian
thalamic artery which often bifurcates off a single pedicle from the posterior
cerebral artery (Castaigne et al., 1981; Von Cramon et al., 1985). LB’s lesions
involved the mammillothalamic and presumably the amygdalofugal tracts, the
intralaminar nuclei, and also parts of the dorsomedial nuclei and ventral lateral
nuclei bilaterally. The ventral anterior nuclei may have also been partially involved.
LB’s initial rCBF brain SPECT scan showed thalamic hyperperfusion compatible
with luxury perfusion following the acute injury (Lassen, 1994). Her subsequent
SPECT scan showed thalamic and basal ganglia hypeperfusion. The basal ganglia
hypoperfusion makes sense with her intralaminar nuclei lesion, since intralaminar
nuclei have major striatal efferent connections (Groenewegen and Berendse, 1994;
Jones, 1985). LB also had general diminished cerebral blood flow and further
reductions in the right frontal and left temporal and parietal lobes.
Patients with bilateral paramedian median thalamic lesions, like LB, typically
have amnesia, decreased levels of arousal and often vertical eye movement
abnormalities (Bogousslavsky et al., 1988b; Guberman and Stuss, 1983;
Tatemichi, Steinke, Duncan et al., 1992). Their amnesia is thought to be a
‘disconnection’ amnesia resulting from deficits in accessing memories that are
stored in a widely distributed fashion (Warrington and Weiskrantz, 1982). The
notion of a ‘disconnection’ amnesia is further supported by neuro-anatomic
observations. These patients invariably have involvement of the mammillothalamic
tract and the ventral amygdalofugal pathways (Crosson, 1992; Gentilini, De
Renzi and Crisi, 1987; Graff-Radford et al., 1990; Von Cramon et al., 1985).
Disruption of these white matter tracts presumably disconnects the frontal lobes
from the ventral-posterior cortical regions. The resulting amnesia is then caused
primarily by a search and retrieval deficit. LB’s discourse suggested a retrieval
problem (discussed below) in keeping with the ‘disconnection’ hypothesis of
diencephalic amnesia. Thus, LB’s amnesia and lesion location are characteristic
of the syndrome of diencephalic amnesia after paramedian thalamic infarcts.
To describe LB as amnestic, while accurate, does not adequately convey the
striking phenomenology of her condition. Her confabulations and often bizarre
responses, mixed with occasional appropriate responses were the dominant
clinical feature. She frequently perseverated, or had intrusions in her speech.
Her output when produced without the imposition of an external structure was
disconnected and had unpredictable shifts to and from fragments of different
themes.
LB performed quite normally on most lexical and morpho-syntactic
parameters of spontaneous speech, despite having a severely disordered speech
output. The proportion of closed class words, noun/pronoun ratio, presence of

432

Anjan Chatterjee and Others

obligatory determiners, noun/verb ratio and frequency of verb inflections in her
speech was normal. She also produced close to normal proportions of words in
sentences and well formed sentences. Her disordered output could not be
accounted for by disruptions of local feature of language production and was
distinct from aphasic disturbances sometimes observed in patients with left
tuberothalamic infarcts (Bogousslavsky et al., 1986b). She also seemed to be
able to access overlearned scripts such as reciting the days of the week, the
alphabet or short rhyming jingles. However, her disorder was most evident in
unstructured discourse.
We think LB’s deficit is better characterized as a dynamic disorder of thought
rather than a language disorder per se. Shanon (1989) argued that thought
sequences may be expressed verbally in ways that offer insight into the workings
of the mind. Other investigators speculate that thalamo-cortical circuit
dysfunction may underlie thought disorders (Crosson and Hughes, 1987; Steriade,
Jones and Llinas, 1990). We are not claiming that LB’s thought disorder is
isomorphic with schizophrenic symptomatology. Rather we wish to underscore
its ideational nature, in which the coherence of underlying streams of thought
is disrupted. Luria (1977) similarly drew attention to an unstable and fluctuating
disorder of speech associated with thalamic lesions, which he felt was not
aphasic, but rather “a disturbance of a deeper nature”.
Confabulations in patients with paramedian thalamic infarcts (Bogousslavsky,
Ferrazzini, Regli et al., 1988a) have been described as resembling ‘psychotic
speech’ (Meissner, Sapir, Kokmen et al., 1987). See Appendix 5 for an example
of LB’s speech that seemed ‘psychotic’. During conversation, LB shifted
thematic topics frequently. The timing of these shifts was unpredictable and the
shifts were usually untethered contextually. At times she would maintain thematic
coherence across several sentences; at other times thematic shifts were evident
within utterances. Elements of more than one theme were sometimes fused in
single utterances. Her inability to maintain or shift thematic topics in a
contextually appropriate manner seems to be a deficit in the regulation of James
‘transitive’ aspects of consciousness. It is reminiscent of Lhermitte’s observation
that these patients have trouble with the ordering of ideas, and passages from one
story seemed to slide into others in their narratives (Lhermitte, 1984). The themes
into which LB unexpectedly shifted were often intrusions from previous topics.
These intrusions were evenly divided into topics that she had introduced and those
introduced by the examiner. Traces of these themes from the conversation had
been encoded and stored sufficiently to resurface (albeit unpredictably) as thematic
intrusions. Yet, at the end of both conversations LB was unable to volitionally
recall any topic discussed. The fact that new information was encoded sufficiently
to later intrude upon her discourse but could not be retrieved volitionally is
compatible with a ‘disconnection’ amnestic state.
LB’s inability to maintain and shift themes appropriately, contrasts with
patient PS, reported by Hodges and McCarthy (1993), who also had bilateral
paramedian thalamic infarcts. PS had significant deficits in learning new material
and had an extensive retrograde autobiographical amnesia. His dominant clinical
feature was the persistent claim that he was home from the navy on leave from
active service during World War II. His memory for early parts of his life was

Thalamic thought disorder

433

fragmentary and could not be integrated into a coherent life story. His retrograde
amnesia was interpreted as a deficit of a ‘thematic retrieval framework’. Hodges
and McCarthy (1993) suggest that episodic memory is retrieved in hierarchically
structured ‘organization packets’. At the highest level, thematic frameworks are
retrieved which provide pointers that guide the selection of lower level detailed
information. The authors suggest that PS was pathologically stuck in the ‘navy’
theme that became a default frame for the reconstruction of his past and
incorporation of new events from his present.
LB’s deficit complements PS’s thematic retrieval deficit. Both LB and PS
were unable to retrieve contextually appropriate themes. PS is tied to a specific
thematic framework in an obligatory manner. By contrast, LB is unable to remain
tethered to a particular framework for any substantial duration. Her deficit results
in unpredictable and rapid shifts in the thematic frameworks guiding her
responses. Many of LB’s ‘confabulatory’ responses might not have been
confabulatory in the specific sense of fictional re-creations (Berlyn, 1972).
Fantastic confabulations, or fictional re-creations, are associated with frontal
dysfunction (Kapur and Coughlan, 1980; Stuss, Alexander, Lieberman et al.,
1978). However, recent reports suggesting that the neural basis for these kinds
of confabulation involves orbito-frontal and medial frontal rather than dorsolateral
prefrontal dysfunction (Benson, Djenderedjian, Miller et al., 1996) support our
notion that LB’s ‘confabulations’ may have been of a different nature. LB’s
peculiar comments may have represented fragmented themes surfacing
unpredictably. From the vantage point of an external observer, without
knowledge of the specific thematic framework guiding her verbal output, she
is confused, confabulatory and at times frankly bizarre. A striking example of
this phenomenon was her drilling her adult daughter on math problems,
something she had regularly done, but over thirty years before her hospital
admission. In keeping with our observations, Shnider and coworkers have recently
reported data to suggest that spontaneous confabulations are associated with an
inability to recognize the temporal order of stored information and reflect a
confusion of memory traces from diverse events rather than a loss of the trace
themselves (Shnider, Daniken and Gutbrod, 1996).
Could LB’s behavior be ascribed to a working memory or executive attention
deficit? She does not have a phonological rehearsal deficit as shown by her
ability to repeat digits and phrases. Her deficit could be due to a specific form
of an executive attentional disturbance along the lines suggested by Norman
and Shallice (1986). According to their model, a large number of thoughts
compete for cognitive resources at any given instance. A ‘supervisory attentional
system’ modulates which thoughts will be activated and which inhibited.
Dysregulation of this system could presumably result in the kind of speech
observed in LB. It remains unclear whether such an attentional disturbance could
be modality specific. Luria thought his similar case with a ‘quasi-aphasic’
disorder had a material specific disorder (Luria, 1977). He suggested that a partial
disturbance of vigilance focused specifically on speech processes was responsible
for his patient’s deficits.
The neural mechanisms underlying LB’s deficit in maintaining and
appropriately shifting themes during spontaneous discourse are not clear. Her

434

Anjan Chatterjee and Others

amnesia probably resulted from the disruption of the mammillothalamic tracts
(Graff-Radford et al., l990; Von Cramon et al., 1985). However, confusion is
not a necessary concomitant of diencephalic amnesia. LB’s confusional state
probably resulted from additional damage to the dorsomedial and intralaminar
nuclei. The dorsomedial nuclei have reciprocal connections with the dorsolateral
prefrontal cortices (Jones, 1985). The intralaminar nuclei have widespread,
topographically organized, cortical reciprocal connections and major striatal
connections (Groenewegen and Berendse, 1994; Jones, 1985). Damage restricted
to the intralaminar nuclei unilaterally may produce clinically significant
distractibility without amnesia (Mennemeier, Fennell, Valenstein et al., 1992).
LB’s dramatic presentation may have been the consequence of a dense amnesia
combined with the widespread effects of intralaminar and midline thalamic
damage.
We do not know if LB’s thalamic lesions produced non-specific cortical
abnormalities or if they contribute to her deficits more directly. Thalamic lesions
may produce cortical diaschisis (Baron, Levasseur, Mazoyer et al., 1992;
Levasseur, Baron, Sette et al., 1992; Pepin and Aurey-Pepin, 1993). Midline
and intralaminar thalamic nuclei are thought to be critical for cortical recruitment
responses (Jones, 1985), perhaps through the regulation of thalamocortical and
striatocortical circuits (Groenewegen and Berendse, 1994). LB’s rCBF brain
SPECT scans showed areas of abnormal cortical perfusion. As such, her deficits
may have been a consequence of abnormally functioning deafferented frontal
and left temporal-parietal cortices. The frontal lobes are conceptualized as critical
for the maintenance of ‘script like’ events, that is, themes encompassing some
temporal duration or regulating the temporal organization of behavior (Fuster,
1989; Grafman, 1994; Shallice and Burgess, 1991). Left temporal-posterior
parietal regions encompass neural networks thought to be critical for the access to
and retrieval of conceptual knowledge (Damasio, Grabowski, Tranel et al., 1996).
It is possible that LB’s dramatically abnormal discourse results from the
abnormality of frontal regulation and maintenance of scripts combined with
problems of access to her conceptual knowledge. As such, her deficits caused
by cortical deafferentation may be more severe and dramatic than deficits
associated with stable lesions restricted to the frontal or left temporal lobes.
An alternate hypothesis posits a more direct thalamic neural mechanism for
LB’s deficit. The very notion of a thematic framework guiding memory searches
and discourse implies highly orchestrated neural activity. The search and
selection of detailed factual information bound to local elements of language,
such as phonological and lexical selection and syntactic structuring, coordinated
with pragmatic aspects of communication must somehow cohere on-line in
discourse. Facial expressions, gestures, body posture somehow yoked to these
themes manifest a coordinated state of consciousness. The problem of
orchestrating neural activity over widely distributed brain regions coordinated
in real time has been termed the binding or the global mapping problem (Llinas
and Pare, 1991). Oscillatory neuronal activity of about 40 Hz are thought to be
a candidate mechanism binding neural activity widely distributed in space within
temporal epochs (Crick and Koch, 1990; Steriade, McCormick and Sejnowski,
1993b). Magnetoencephalographic recordings in awake humans reveal the

Thalamic thought disorder

435

presence of continuous 40 Hz oscillations over the entire cortical mantle (Llinas
and Ribary, 1993). Thalamo-cortical circuits may regulate such oscillations over
distributed cortical regions via diffuse cortical projections with reentrant back
projections (Steriade et al., 1990). Intralaminar nuclei neurons generate 40 Hz
oscillations (Steriade, Dossi and Contreras, 1993a). These specific oscillations
are postulated to play a critical role in the binding necessary for the unity of
cognitive experience (Crick and Koch, 1990; Llinas and Ribary, 1993; Newman,
1995). Perhaps damage to the intralaminar nuclei impairs the binding of neural
activity over widely disparate neural structure. In its mild form, this impairment
might result in increased distractibility (Mennemeier et al., 1992). In its severe
form, perhaps combined with amnesia, this impairment might result in an
obligatory entrapment within a specific thematic framework (Hodges and
McCarthy, 1993), or in the untethering of thematic frames from volitional control
or contextual constraints.
In conclusion, we think that LB’s confusion is not explained by disruptions
of local features of language. We echo Luria’s suggestion (from 20 years ago)
that this dramatically disordered speech associated with thalamic lesions is the
surface manifestation of a deeper dynamic disorder. This disorder is characterized
by thematic instability and fragmentation. Damage to midline thalamic nuclei
may have widespread cortical effects resulting in the inability to maintain and
appropriately shift themes within discourse. The instability of thematic
frameworks results in frequent and unpredictable thematic shifts as evidenced
by abnormalities of the “rhythm of language”.
Acknowledgements. We thank Lisa Santer for a critical reading of the manuscript and
Angela Armstrong for discourse transcriptions.

REFERENCES
BARON, J., LEVASSEUR, M., MAZOYER, B., LEGAULT-DEMARE, F., MAUGUIERE, F., PAPPATA, S.,
JEDYNAK, P., DEROME, P., CAMBIER, J., TRAN-DINH, S., and CAMBON, H. Thalamocortical diascisis:
positron emission tomography in humans. Journal of Neurology, Neurosurgery and Psychiatry, 55:
935-942, 1992.
BENSON, D.F., DJENDEREDJIAN, A., MILLER, B.L., PACHANA, N.A., CHANG, L., ITTI, L., ENG, G.E.,
and MENA, I. Neural basis of confabulation. Neurology, 46: 1239-1243, 1996.
BERLYN, N. Confabulation. British Journal of Psychiatry, 120: 31-39, 1972.
BERNDT, R.S. Sentence processing in aphasia. In M. Sarno (Eds.), Acquired Aphasia. New York:
Academic Press, 1991, pp. 223-269.
BLOOM, R.L. Hemispheric responsibility and discourse production: contrasting patients with unilateral
left and right brain damage. In R.L. Bloom, L.K. Obler, S.D. Santi and J.S. Ehrlich (Eds.), Discourse
Analysis and Applications. Hillsdale: Lawrence Erlbaum Associates, 1994, pp. 81-94.
BOGOUSSLAVSKY, J., FERRAZZINI, M., REGLI, F., ASSAL, G., TANABE, H., and DELALOYE-BISCHOF, A.
Manic delirium and frontal-like syndrome with paramedian infarction of the right thalamus. Journal
of Neurology, Neurosurgery, and Psychiatry, 51: 116-119, 1988a.
BOGOUSSLAVSKY, J., REGLI, F., and USKE, A. Thalamic infarcts: Clinical syndromes, etiology, and
prognosis. Neurology, 38: 837-848, 1988b.
CASTAIGNE, P., LHERMITTE, F., BUGE, A., ESCOUROLLE, R., HAUW, J.J., and LYON-CAEN, O. Paramedian
thalamic and midbrain infarcts: Clinical and neuropathological study. Annals of Neurology, 10:
127-148, 1981.
CHANG, L.T. A method for attenuation correction in radionuclide computed tomography. IEEE
Transactions of Nuclear Sciences, 26: 638-643, 1978.
CHATTERJEE, A., and MENNEMEIER, M. Anosognosia for hemiplegia: patient retrospections. Cognitive
Neuropsychiatry, 1: 221-237, 1996.

436

Anjan Chatterjee and Others

CRICK, F., and KOCH, C. Towards a neurobiological theory of consciousness. Seminars Neuroscience,
2: 263-275, 1990.
CROSSON, B. Subcortical Functions in Language and Memory. New York: The Guilford Press, 1992.
CROSSON, B., and HUGHES, C. Role of the thalamus in language: is it related to schizophrenic thought
disorder? Schizophrenia Bulletin, 13: 605-621, 1987.
CROW, B. Topic shifts in couples’ conversation. In R. Craig and K. Tracy (Eds.), Conversational
Coherence: Form, Structure and Strategy. Beverly Hills, CA: Sage, 1983.
DAMASIO, H., GRABOWSKI, T., TRANEL, D., HICHWA, R.D., and DAMASIO, A.R. A neural basis for
lexical retrieval. Nature, 380: 499-505, 1996.
DELIS, D., KRAMER, J., KAPLAN, E., and OBER, B. The California Verbal Learning Test Adult Version:
Manual. New York: Psychological Corporation, 1987.
FUSTER, J. The Prefrontal Cortex (2nd ed.). New York: Raven Press, 1989.
GARCIA, L.J., and JOANETTE, Y. Conversational topic-shifting analysis in dementia. In R.L. Bloom,
L.K. Obler, S. DeSanti and J.S. Ehrlich (Eds.), Discourse Analysis and Applications. Hillsdale,
New Jersey: Lawrence Erlbaum Associates, 1994.
GENTILINI, M., DE RENZI, E., and CRISI, G. Bilateral paramedian thalamic artery infarcts: report of
eight cases. Journal of Neurology, Neurosurgery, and Psychiatry, 50: 900-909, 1987.
GRAFF-RADFORD, N.R., TRANEL, D., VAN HOESEN, G.W., and BRANDT, J.P. Diencephalic amnesia.
Brain, 113: 1-25, 1990.
GRAFMAN, J. Alternative frameworks for the conceptualization or prefrontal lobe functions. In F. Boller
and J. Grafman (Eds.), Handbook of Neuropsychology, Vol. 9. Amsterdam: Elsevier Science
Publishers, 1994, pp. 187-202.
GROENEWEGEN, H.J., and BERENDSE, H.W. The specificity of the ‘nonspecific’ midline and intralaminar
thalamic nuclei. Trends Neuroscience, 17: 52-58, 1994.
GUBERMAN, A., and STUSS, D. The syndrome of bilateral paramedian thalamic infarction. Neurology,
33: 540-546, 1983.
HAINES, D. Neuroanatomy. An Atlas of Structure, Section, and Systems (4rth ed.). Baltimore: Williams
and Wilkins, 1995.
HODGES, J.R., and MCCARTHY, R.A. Autobiographical amnesia resulting from bilateral paramedian
thalamic infarction: A case study in cognitive neurobiology. Brain, 116: 921-940, 1993.
JAMES, W. The Principles of Psychology. New York: Dover Publications, Inc., 1890.
JONES, E. The Thalamus. New York: Plenum Press, 1985.
KAPUR, N., and COUGHLAN, A.K. Confabulation and frontal lobe dysfunction. Journal of Neurology,
Neurosurgery, and Psychiatry, 43: 461-463, 1980.
LASSEN, N.A. Imaging brain infarcts by single-photon emission tomography with new tracers. European
Journal of Nuclear Medicine, 21: 189-190, 1994.
LEVASSEUR, M., BARON, J.C., SETTE, G., LEGAULT-DEMARE, F., PAPPATA, S., MAUGUIERE, F., BENOIT,
N., TRAN-DINH, S., DEGOS, J., LAPLANE, D., and MAZOYER, B. Brain energy metabolism in bilateral
paramedian thalamic infarcts. Brain, 115: 795-807, 1992.
LHERMITTE, F. Language disorders and their relationship to thalamic lesions. Advances in Neurology,
42: 99-113, 1984.
LIU, H.G., HARRIS, J.M., INAMPUDI, C.S., and MOUNTZ, J.M. Optimal reconstruction filter parameters
for multi-head brain SPECT: dependence on count activity. Journal of Nuclear Medicine
Technology, 23: 251-257, 1995.
LLINAS, R., and PARE, D. Of dreaming and wakefulness. Neuroscience, 44: 521-535, 1991.
LLINAS, R., and RIBARY, U. Coherent 40-Hz oscillation characterized dream states in humans.
Proceedings of National Academy of Sciences, 90: 2078-2081, 1993.
LURIA, A.R. On quasi aphasic speech disturbances in lesions of deep structures of the brain. Brain and
Language, 4: 432-459, 1977.
MATTIS, S. Mental status examination for organic mental syndrome in the elderly patient. In L. Bellak
and T.B. Karasu (Eds.), Geriatric Psychiatry: New York: Grune and Stratton, 1976.
MEISSNER, I., SAPIR, S., KOKMEN, E., and STEIN, S.D. The paramedian diencephalic syndrome: A
dynamic phenomenon. Stroke, 18: 380-385, 1987.
MENNEMEIER, M., FENNELL, E., VALENSTEIN, E., and HEILMAN, K. Contributions of the left intralaminar
and medial thalamic nuclei to memory. Archives of Neurology, 49: 1050-1058, 1992.
MILNER, A.D., and RUGG, M.D. The Neuropsychology of Consciousness. New York: Academic Press,
1992.
MOUNTZ, J.M. Quantification of the SPECT brain scan. In L.M. Freedman (Eds.), Nuclear Medicine
Annual. New York: Raven Press, 1991, pp. 67-98.
MOUNTZ, J.M. Brain SPECT: 1994 Update. In L.M. Freeman (Ed.), Nuclear Medicine Annual. New
York: Raven Press, 1994, pp. 1-54.
MOUNTZ, J.M., WILSON, M.W., WOLFF, C.G., DEUTSCH, G., and HARRIS, J.M. Validation of a reference
method for correlation of anatomic and functional brain images. Computerized Medical Imaging
Graphics, 18: 163-174, 1994a.

Thalamic thought disorder

437

MOUNTZ, J.M., ZHANG, B., LIU, H.-G., and INAMPUDI, C. A reference method for correlation of anatomic
and functional brain images: Validation and clinical application. Seminars in Nuclear Medicine,
24: 256-271, 1994b.
NEWMAN, J. Thalamic contributions to attention and consciousness. Consciousness and Cognition, 4:
172-193, 1995.
NORMAN, D., and SHALLICE, T. Attention to action: willed and automatic control of behavior. In R.J.
Davidson, G.E. Schwartz and D. Shapiro (Eds.), Consciousness and Self-Regulation. New York:
Plenum Press, 1986.
PEPIN, E., and AUREY-PEPIN, L. Selective dorsolateral frontal lobe dysfunction associated with
diencephalic amnesia. Neurology, 43: 733-741, 1993.
RUSSELL, E.W. A multiple scoring method for the assessment of complex memory functions. Journal
of Consulting Clinical Psychology, 43: 800-809, 1975.
SAFFRAN, E.M., BERNDT, R.S., and SCHWARTZ, M.F. The quantitative analysis of agrammatic production:
procedure and data. Brain and Language, 37: 440-479, 1989.
SHALLICE, T., and BURGESS, P. Higher-order cognitive impairments and frontal lobe lesions. In H.
Levin, H. Eisenberg and A. Benton (Eds.), Frontal Lobe Function and Dysfunction. New York:
Oxford University Press, 1991, pp. 125-138.
SHANON, B. Thought sequences. European Journal of Cognitive Psychology, 1: 129-159, 1989.
SHNIDER, A., DANIKEN, C.V., and GUTBROD, J. The mechanisms of spontaneous and provoked
confabulations. Brain, 119: 1365-1375, 1996.
STERIADE, M., DOSSE, R.C., and CONTRERAS, D. Electrophysiologic properties of intralaminar
thalamocortical cells discharging rhythmic (40Hz) spike-bursts at 1000 Hz during waking and
rapid eye movement sleep. Neuroscience, 56: 1-9, 1993a.
STERIADE, M., JONES, E., and LLINAS, R. Thalamic Oscillations and Signalling. New York: John Wiley
and Sons, 1990.
STERIADE, M., MCCORMICK, D., and SEJNOWSKI, T.J. Thalamocortical oscillations in the sleeping and
aroused brain. Science, 262: 679-685, 1993b.
STUSS, D.T., ALEXANDER, M.P., LIEBERMAN, A., and LEVINE, H. An extraordinary form of confabulation.
Neurology, 28: 1166-1172, 1978.
TATEMICHI, T.K., STEINKE, W., DUNCAN, C., BELLO, J.A., ODEL, J.G., BEHRENS, M.M., HILAL, S.K.,
and MOHR, J.P. Paramedian thalamopeduncular infarction: Clinical syndromes and magnetic
resonance imaging. Annals of Neurology, 32: 162-171, 1992.
VON CRAMON, D.Y., HEBEL, N., and SCHURI, U. A contribution to the anatomical basis of thalamic
amnesia. Brain, 108: 993-1008, 1985.
WARRINGTON, E.K., and WEISKRANTZ, L. Amnesia: A disconnection syndrome. Neuropsychologia, 20:
233-249, 1982.
Anjan Chatterjee, UAB Neurology, 1720 7th Ave South, SC454, Birmingham, AL 35294-0017, USA. Tel.: (205) 934-3847; Fax:
(205) 975-7365; email: a_chatterjee@email.neuro.uab.edu.

(Received 18 September 1996; accepted 5 March 1997)

438

Anjan Chatterjee and Others

APPENDIX
LB’s most striking clinical feature was her apparent confusion and bizarre responses. To
convey the quality of her phenomenology we have included examples of extended excerpts.
Examples of coding are included at the beginning of the shift, following the abbreviations
listed in Table II.
(1) LB has been told that she was being videotaped for research purposes. We think that
this information served as the seed for ‘research’ appearing in her subsequent responses. We
do not know the context for the girl scouts’ theme. However, towards the end of this excerpt
both research and girls scouts were merged into a single utterance.
E:
LB:
E:
LB:
E:
LB:
E:
LB:
E:
LB:
E:
LB:
E:
LB:
E:
LB:
E:
LB:

Do you know how long you have been here?
I’ve been here about two days, /1.R, 2.U, 3.ET/ doing a research project for another
group of girls.
What kind of research project are you doing?
This girls scout project, with... I was a background project.
Do you think this a hospital?
No. No, this is... /1.U, 2.U, 3.U/ You don’t do any more ground work /1.U, 2.U, 3.U/
I suppose we do in learning about people.
Have you been ill at all?
No. The only illness I have had was diagnosed by your company as a result of /1.U,
2.U, 3.U/ setting on duty for around here, for an over extended length of time, when
I was examined. /1.U, 2.U, 3.U/ I think yesterday you had... as a result of...
How’s your thinking?
It’s been a lot better. /1.U, 2.U, 3.U/ I influenced from... and /1.R, 2.U, 3.ET/ our
research has been outdoor work of bees and butter... interior.
Where are you from?
I am from Texas.
Did you grow up there?
Yes.
How did you come to Alabama from Texas?
As a research scientist.
What were you doing research in?
/1.C, 2.U, 3.ET/ My sister was in research with the girl scouts and I just kind of
followed in her tracks.

(2) Sometimes LB simply confabulated responses. The source of these confabulations
could not be determined from the context of the conversation or any information available
about LB.
E:
LB:
E:
LB:
E:
LB:

Do you know where you are? Where is this place?
South America.
South America?
/1.TD, 2.U, 3.ET/ S-O-U-T-H, That’s the way to spell South America.
Why are you in South America?
My husband. He’s in South America. He’s supposed to be there tonight.

(3) At times, LB became ‘stuck in set’ and would thematically perseverate as the following
excerpt demonstrates. We elected to count such perseverations as topic shifts since LB shifts
to her stuck-in-set topic rather than the new topic introduced be the examiner.
E:
LB:
E:
LB:
E:
LB:
E:

Can you name for me as many of the states of the US that you can think of?
Alabama, Mississippi, Georgia, Kentucky...
Keep going.
Kentucky, Wyoming. (prolonged pause, stays quiet).
Ma’am?
Yes?
What are you thinking of?

Thalamic thought disorder

LB:
E:
LB:
E:
LB:
E:
LB:
E:
LB:
E:
LB:

439

Alabama, Mississippi, Georgia, Chicago. Chicago is... Mississippi and Georgia are
states. Georgia is. Mississippi, Alabama, they’re towns (another long pause).
What is the matter?
/1.R, 2.U, 3.ET/ Alabama, Georgia. (long pause).
Ma’am, can you describe your house for me?
Which house?
The house you live in.
/1.R, 2.U, 3.ET/ Alabama, Georgia, Maine, Florida. (long pause).
I’d like you to try something else. Can you name as many animals as you can think
of?
/1.R, 2.U, 3.ET/ Alabama, Georgia, Mississippi.
Ma’am, can you name as many animals as you can think of?
Fox, rabbit.

(4) Attempts to get LB to narrate fairy tails were unsuccessful. She says enough
(Cinderella-attic, sisters; Red Riding Hood-woods) to show that she made the correct
association. However, a ‘home’ theme that was present at different times in her discourse
intrudes. The peculiar numbering from 11 to 13 associated with Christmas persists across
attempts at both fairy tales.
E:
LB:
E:
LB:
E:
LB:
E:
LB:
E:
LB:
E:
LB:

E:
LB:

Do you remember the Cinderella story?
Home up in the attic.
That’s right. She had a home up in the attic.
She didn’t get along with her sisters. /1.R, 2.US 3.ET/ Be home before Christmas. 11
things like Christmas. Find 11 things like Christmas.
What about 11 things like Christmas?
/1.R, 2.U, 3.ET/ She was at home. /1.C, 2.U, 3.ET/ Her home was 12 things for
Christmas.
What else happened in the Cinderella story? (non answer)
Did she go to a dance?
She went to... /1.U, 2.U, 3.U/ store.
Did she go to a dance? A ball?
A ball is a big dance. /1.R, 2.U 3.IT/ Home.
Do you remember Red Riding Hood?
She was out in the woods, but, she... /1.R, 2.U, 3.IT/ I don’t know anything about
her home, her parents. Oh, yes. /1.R, 2.U, 3.IT/ 13 things. There are 13 people, 13
things to remember. 13 things at home. There were 13 people who lived where she
lived.
Where who lived?
/1.C, 2.U, 3.ET/ 13 different Christmases.

(5) LB’s discourse, at time, was sufficiently bizarre as to fit previous descriptions of
similar patients having ‘psychotic speech’.
E:
LB:
E:
LB:
E:
LB:

Can you tell me the Cinderella story?
Not tonight.
How about another fairy tale? Can you tell me Sleeping Beauty?
Don’t they all have the same story? Are they... /1.R, 2.U, 3.ET/ fairy godmother who
would...
Who would do what?
Exploit the boy with tales or stories that he would not believe were they told in any
other context. /1.U, 2.U, 3.U/ I can’t see him. /1.U, 2.U, 3.U/ He would have to
experience... There aren’t any more mosquitoes that serve in the civilian guard. Serve.
/1.R, 2.U, 3.ET/ For him to believe it, and so with all stories.

(6) Sometimes a question or word seemed to trigger a train of thought in LB, only loosely
related to the initial suggestion. In the following excerpt, LB remains within the general
bounds of her theme for over three minutes, the longest period during the time she was

440

Anjan Chatterjee and Others

recorded. She was a deeply religious woman. In addition to the poetry of her output, what
is not easily conveyed in the following excerpt is the cadence, prosody and modulation of
volume reminiscent of a minister giving a sermon.
E:
LB:
E:
LB:
E:
LB:

Do you know what month this is ma’am?
We’re in February.
Actually it is December. Can you think of what holiday happens in December?
Oh! December, you know. /1.TD, 2.U, 3.ET/ December is our greatest month, is our
greatest gift. Our greatest prophet is supposed to have come in December and given
us our greatest gift, a child who was like nothing else in this world.
How do we celebrate that day?
We call it Christmas and we give gifts. We know what is in people’s heart is true or
is false. Just December means everything is perfect. December means everybody is
good. December means I love you and you love me. December means all is good,
all is perfect, all is God. December means what that sound means, all is well, all will
move in the right direction, all is wise, all is wisdom, all is light, all is truth, all is
goodness. December means there is no separation one from another. December means
you come through me and learn all you need to learn and I go through you and learn
all I need to know. December means God is open for all his people to use. God is
there for all his people to use. God is one for all his people to use. God is perfect.
He’s whatever number we need. God is perfect. He’s perfect. He’s whatever number
we need. God is perfect.