Author’s Accepted Manuscript
Testing the importance of the Medial Temporal
Lobes in human interoception: Does it matter if
there is a memory component to the task?
Joanne Berriman, Richard J. Stevenson, Zoe C.
Thayer, Elizabeth Thompson, Armin Mohamed,
John D.G. Watson, Laurie A. Miller
www.elsevier.com/locate/neuropsychologia

PII:
DOI:
Reference:

S0028-3932(16)30342-6
http://dx.doi.org/10.1016/j.neuropsychologia.2016.09.005
NSY6127

To appear in: Neuropsychologia
Received date: 28 April 2016
Revised date: 23 August 2016
Accepted date: 4 September 2016
Cite this article as: Joanne Berriman, Richard J. Stevenson, Zoe C. Thayer,
Elizabeth Thompson, Armin Mohamed, John D.G. Watson and Laurie A. Miller,
Testing the importance of the Medial Temporal Lobes in human interoception:
Does it matter if there is a memory component to the task?, Neuropsychologia,
http://dx.doi.org/10.1016/j.neuropsychologia.2016.09.005
This is a PDF file of an unedited manuscript that has been accepted for
publication. As a service to our customers we are providing this early version of
the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting galley proof before it is published in its final citable form.
Please note that during the production process errors may be discovered which
could affect the content, and all legal disclaimers that apply to the journal pertain.

Testing the importance of the Medial Temporal Lobes in human interoception: Does it
matter if there is a memory component to the task?
Joanne Berrimana, Richard J. Stevensona*, Zoe C. Thayerb, e, Elizabeth Thompsonc, Armin
Mohamedd, e, f, John D.G. Watsonf, Laurie A. Millerb, e,
a

Department of Psychology, Macquarie University, Sydney, NSW2109, Australia.

b

c

Neuropsychology Unit, Royal Prince Alfred Hospital, Sydney, NSW2050, Australia.

Department of Radiology, Royal Prince Alfred Hospital, Sydney, NSW2050, Australia.

d

Neurology Department, Royal Prince Alfred Hospital, Sydney, NSW2050, Australia

e

ARC Centre of Excellence in Cognition and its Disorders, University of Sydney, Sydney,

NSW2006, Australia.
f

Central Medical School, University of Sydney, Sydney, NSW2006, Sydney, Australia.

g

Neurology Department, Sydney Adventist Hospital, Sydney, NSW2076. Sydney, Australia.

*

Communicating author: R.J. Stevenson, Department of Psychology, Macquarie University,

Sydney, NSW2109, Australia dick.stevenson@mq.edu.au phone 61-2-98508098; fax 61-298508062
Abstract
Interoception is the ability to consciously perceive internal bodily states. Neuroimaging
suggests that the insula (IC) and anterior cingulate cortex (ACC) mediate interoception, while
studies involving patients/animals with brain lesions suggest the medial temporal lobe (MTL)
is particularly important. One reason for these contrasting conclusions may lie in the types of
interoceptive task used by these different approaches. Some tasks probably require integration
of current physiological state with mnemonic information (e.g., how much one last ate), and
these may be especially reliant upon MTL processing. We compared one task that probably
requires integration - a water load task - with one that likely does not - a heart-rate tracking
task - in two individuals with selective MTL damage (and with intact IC and ACC). A group

2

of matched healthy individuals served as controls. The main finding was that individuals with
MTL damage, relative to controls, were equally and significantly impaired on both types of
interoception task. This suggests that MTL structures are involved in mediating interoception
even when using a task (heart rate tracking) that does not seemingly require memory and that
in neuroimaging studies activates the IC and ACC. The reasons for this apparent inconsistency
with neuroimaging findings and the functional role of the MTL in interoception are discussed.

Key Words: Bodily states, Heart-rate tracking, Hippocampus, Insular Cortex, Memory,
Appetite

1. Introduction
Interoception is the ability to perceive internal bodily states such as heartbeat, hunger, pain
and fullness (Craig, 2002; Pollatos et al., 2007a; Van Oudenhove et al., 2004). Human studies
suggest that medial temporal lobe (MTL) lesions can generate interoceptive impairments
(Hebben et al., 1985; Higgs et al., 2008; Rozin et al., 1998). Anatomically, the MTL includes
three major parts: (1) the olfactory system; (2) the hippocampus, entorhinal, perirhinal and
parahippocampal cortices; and (3) the amygdala - with the latter structures (notably (2))
supporting declarative memory (Squire, Stark & Clark, 2004). The MTL’s putative role in
interoception was first suggested by case HM. After his bilateral MTL resection HM rarely
reported being hungry, thirsty or tired, and had a diminished capacity to feel pain (Hebben et
al., 1985). What makes the results from HM and other MTL cases so puzzling, is that
neuroimaging studies typically suggest that the insula cortex (IC) and the anterior cingulate
cortex (ACC) are crucial for mediating interoception (Craig, 2002, 2003, 2009). Yet HM,
whose brain has now been dissected, appears to have had no damage to his IC or ACC (Annese
et al., 2014). In this study we explore one possible explanation for these divergent conclusions

3

and study the extent to which selective MTL lesions, sparing the IC and ACC, affect
interoception.
Neuroimaging research, mainly using PET and fMRI, has employed a range of tasks to
investigate the neural correlates of interoception. These have included: oesophageal
stimulation (e.g., Pollatos et al., 2007a); gastric distension (e.g., Wang et al., 2008); respiratory
tasks (Rosenkranz et al., 2005); rectal stimulation (Berman et al., 2008); sexual responses
(Komisaruk et al., 2004); bladder filling (Mehnert et al., 2008); thirst (Denton, et al., 1999);
and heart rate tracking (Pollatos et al., 2007b). In line with a theory advanced by Craig (2002,
2003, 2009), which states that the IC has a primary role (and to a lesser extent the ACC) in
supporting introspection, every study examined across these various task-types reported
activation of the IC, with around half also reporting activation of the ACC. MTL activation
was only more rarely observed.
Neuropsychological studies of interoception started with HM, finding that he was unable
to process information related to the sensations of hunger, thirst, fullness and pain (Hebben et
al., 1985). As noted above, HM had an almost complete removal of the MTL (i.e., entorhinal
cortex, anterior hippocampus, and most of the amygdala in both hemispheres) but with no
lesions to his IC and ACC (Annese et al., 2013). The importance of the MTL in mediating
interoception has been confirmed in two subsequent studies of patients with bilateral MTL
lesions. All reported interoceptive deficits in hunger and fullness – the sole modalities
explored in these studies (Higgs et al., 2008; Rozin et al., 1998). The importance of MTL
structures in interoceptive processing is also suggested by two further sets of findings. The
first concerns animal studies, with selective lesions of the hippocampus resulting in
interoceptive deficits in hunger and satiety (Davidson and Jarrad, 1993; Davidson et al., 2009)
and dysregulation of food intake (Clifton et al., 1998). The second concerns neuroimaging.
While nearly all studies of gastric interoception demonstrate insula involvement, some also

4

reveal MTL activation as well (including the amygdala), a finding which may be particular to
ingestion-related tasks (e.g., DelParigi et al., 2004; Geeraerts et al., 2011).
Our contention is that certain types of interoceptive task may draw more heavily on MTL
structures than others. In particular, a key difference may lie in whether the task requires
integration of information, namely between current physiological state and past behaviour
(e.g., for fullness – what is felt now with recollection of what was eaten before). The literature
on effects of MTL lesions on interoception in humans and animals has focussed mainly on
eating related variables. This type of variable may be just the sort of task that is influenced by
both current visceral state (e.g., feelings of hunger) and by the memory of past events (e.g.,
food ingested), requiring the integration of these two pieces of information – a task well suited
to the MTL, and especially the hippocampus. One reason to suggest this possibility comes
from a study by Brunstrom et al., (2012), where both how much food participants actually ate
and how much they thought they ate, were manipulated. They found that fullness ratings
shortly after ingestion (i.e., current physiological state) were primarily influenced by what had
actually been eaten, but with time, what the participants thought they had eaten became the
more dominant influence (i.e., integration of memories of food intake with physiological state
data). While the latter task may require MTL involvement, it is possible that tasks that are
reliant upon just current physiological state alone, would not. Thus, in this case, we would
expect MTL lesions to have no effect.
In the current study we therefore included two types of interoceptive task, one that we
thought would involve the MTL and one that would not – respectively - The Water Load Task
(WLT) and the Heart-Rate Tracking (HRT) task. The WLT involves rating sensations of
fullness prior to drinking water until feeling comfortably full, and then making further fullness
ratings at two later time points (based upon, Herbert et al., 2012). The HRT task requires
participants to count their heartbeat and is a good example of a task that probably does not

5
involve any memory integration – just monitoring of current physiological state. Indeed, the
HRT task is the most suitable measure for just assessing cardiac interoception, as its main
alternative the heart beat detection task assesses multisensory integration of interoceptive and
exteroceptive information (Michal et al., 2014; Schulz et al., 2013). One potential concern
with the HRT task is that participants who may be unable to discern their heart beat could then
try to use explicit knowledge of their heart rate and use this to mentally calculate a response
(see Ring et al., 2015). To forestall this possibility, we instructed participants that if they were
completely unable to discern their heart beat they were to report being unable to do so on that
trial. Consistent with many other interoceptive tasks, HRT has been shown to activate the IC
and ACC in fMRI studies, but not MTL structures (Pollatos et al., 2007b).
As we planned to test patients with MTL lesions - relative to healthy matched controls and as MTL structures (especially the hippocampus) are critical for episodic memory (e.g.,
Jeneson & Squire, 2012), measures of immediate and delayed memory (Logical Memory;
WAIS-IV, Wechsler, 2008) were administered to confirm the expected pattern of deficit.
Measures of working memory (Digit Span; WAIS-IV, Wechsler, 2008) were also included, as
this variable was not expected to differ between patients and controls, as working memory is
not dependent upon the MTL (e.g., Canavan et al., 1989). Two control tasks (The Water Level
test and the Tone Counting task) were also included. These mirrored the non-interoceptive
aspects of the HRT task and WLT involving making judgements and counts, so as to look for
any such underlying task-specific deficits. Finally, variability in interoceptive performance is
associated with anxiety (Pollatos et al., 2007a), fitness (Vaitl, 1996), physical symptom
reporting (Duddu et al., 2003), depression (Pollatos et al., 2009), obesity (Herbert & Pollatos,
2014), and sleep deprivation (Greer et al., 2013). All of these variables were measured in order
to identify any relevant confounds between patients and controls.

6

In sum, we hypothesised that individuals with selective damage to the MTL and with no
damage to the IC or ACC, would demonstrate interoceptive deficits on the WLT, as this likely
involves the integration of gastric visceral sensations with memory of prior water intake. In
contrast, it was hypothesised that there would be no difference between the patients and
controls on the HRT task, as this does not require integration. Thus, we predicted an
interaction between task type and ‘group’ (i.e., patients vs. healthy controls).

2. Material and Methods
2.1 Participants
Two patients were recruited from the Neuropsychology Unit, Royal Prince Alfred Hospital
(RPAH), Sydney. PT, was a 40 year old right-handed female, currently employed as a bank
loan-manager, who underwent a right temporal lobectomy in 2002 for intractable epilepsy. PT
reported memory problems, with a neuropsychological assessment conducted at RPAH (one
year prior to testing here) indicating deficits in memory for both visual and verbal material. PT
reported feeling no sensations of hunger but experienced feelings of fullness after eating. She
reported that she often forgot to drink, did not experience thirst, and would only remember to
drink if she developed a headache due to dehydration.
An MRI conducted six months prior to testing confirmed a partial right temporal
lobectomy with resection of the head and body of the right hippocampus as well as
encephalomalacia in the residual temporal lobe involving the superior temporal gyrus on the
right (see Fig.1a). No other abnormalities were reported.

The second patient, JW, was a 62 year old left-handed female retired book-keeper, who
presented with a six month history of memory loss. Neuropsychological assessment conducted

7

at RPAH (3 months prior to testing for this project) detected marked impairments on tests of
verbal and visual anterograde memory. JW also reported recent weight loss (1-2 kg in the last
two weeks), which was associated with a lack of appetite, and forgetting to eat. This was of
concern to her carer, who had taken to reminding her to eat throughout the day. JW reported
drinking regularly without reminders.
Magnetic resonance imaging (MRI) was conducted three months prior to the study (see
Fig.1b). The imaging revealed symmetrical, bilateral atrophy of both hippocampi with
enlargement of the temporal horns, suggestive of mesial temporal sclerosis. Volume loss was
also evident in the right amygdala. In addition, there was a small lacune in the white matter of
the right parietal lobe lying just above the occipital horn of the right lateral ventricle.
Elsewhere, there was evidence of white matter hyperintensities in cerebral hemispheric white
matter consistent with lesions due to small vessel ischaemia.
Thirty healthy controls participated in the study - 5 males (17%) and 25 females (83%).
These included mature-aged students recruited from Macquarie University (n = 13) and people
from the local community (n = 17). Inclusion criteria were: (1) age range of 30 to 65 years,
and (2) English as a first language. Exclusion criteria were based upon factors that could affect
task performance, influence interoception or affect neuropsychological testing: (1)
hospitalization within the last six months; (2) a history of any brain disease or trauma which
could potentially impair memory; (3) medical conditions (e.g., diabetes) and/or substance
abuse that could impair cognitive functioning; (4) psychiatric illness; (5) bladder or kidney
dysfunction; (6) prescription medications that affect attention/concentration; and (7) a history
of eating disorders.
The study was approved by Macquarie University Human Research Ethics Committee and
by the RPAH Ethics Committee. Informed consent was provided by all participants.
2.2 Materials

8

2.2.1. Biographical and background questionnaire data
Biographical and medical history was obtained from each participant, as well as
measuring their weight and height so as to calculate Body Mass Index (BMI; weight in kg
divided by height in M squared). Three questionnaires, with established reliability and
validity, were also completed: (1) The Pittsburgh Insomnia Rating Scale (PIRS-2; Moul et al.,
2002); (2) The International Physical Activity Questionnaire – short form (IPAQ; Craig et al.,
2003); and (3) The Depression Anxiety Stress Scale (DASS-21; Lovibond & Lovibond, 1995).
Respectively, higher scores on these three scales indicate poorer sleep, higher physical activity
levels, and greater depression, stress and anxiety.
2.2.2 Interoception tasks
Interoceptive sensitivity was assessed with a version of the heart rate tracking (HRT) task
(Herbert et al., 2007, 2010; Schandry, 1981). After an initial rest period of 60 seconds,
participants were asked to concentrate on their heart, and to count their heartbeats over four
counterbalanced time intervals of 25, 35, 45, and 55 seconds. Each counting period was
separated by a standard resting period of 30 seconds, and a start and stop cue was given at the
beginning and end of the counting period. At the end of each period participants were asked to
verbally report the number of heartbeats counted, unless they were completely unable to
discern their own heart beating. Participants were instructed to abstain from manually taking
their pulse or any similar direct means of assessing heart rate. Participants were blind to the
length of the counting phase and were not provided with feedback on their performance.
During each trial, participants’ heart rate was monitored with a Pulse Oximeter (Crucial
Medical Systems). This test was conducted twice to confirm its reliability. After the
conclusion of the second test, participants were asked to report how they went about the task.
An absolute difference was obtained by subtracting the estimated heart rate from the actual
heart rate for each of the four time periods on the first and then the repeat test. In line with

9

Schandry (1981), each absolute difference score was then divided by actual heart rate, and the
value subtracted from one. A score of 1 indicates perfect interoceptive ability on the HRT task,
with scores approaching 0 indicating poor performance. Group means were calculated on these
values for each of the two HRT tasks, providing two absolute value scores. These were
collapsed into a single value (control range 0.04 to 0.93) after reliability (for patients and
controls) was verified using the Intraclass Correlation Coefficient, r = .95. Two controls were
excluded from this analysis. One was unable to sense their heart rate on any trial and one
managed to do so on only one trial. Of the remaining controls, four were unable to sense their
heart rate on certain trials - two on three trials and two on one trial. Both, patients reported
being unable to sense their heart rate on particular trials, JW failed to provide a response on
one occasion and PT provided responses on just two trials. In these cases of missing trials, the
overall score was based on the remaining trials.
The water load task (WLT) was based upon Herbert et al., (2012). Participants were
presented with a clear, five litre canister containing three litres of room temperature, noncarbonated water. They were asked to drink as much as they required over a five min period
until they felt comfortably full. Identical 150 ml cups were used for each participant. The
amount of water was measured prior to the participant arriving and after they had departed so
as to calculate the volume consumed during the drinking phase. A key component of the WLT
involved each participant measuring their subjective feelings of fullness on a 15cm visual
analogue (VA) scale (anchors: ‘Not at all’ to ‘Very’) over three time periods: 15 minutes prior
to drinking, immediately after drinking, and 30 minutes later. So as not to focus attention
solely on fullness, participants’ also rated hunger and thirst using similar scales.
Linear and quadratic slope coefficients were calculated for the three fullness ratings (i.e.,
across the three time periods) on the WLT. Positive linear slopes indicated increasing fullness
over time, negative values, decreasing. Positive quadratic coefficients indicated falling and

10

then rising fullness over time, negative quadratic coefficients indicated increasing and then
decreasing fullness over time.
We also included a self-report measure that assesses habitual attentiveness to bodily
sensations – the Pennebaker Inventory of Limbic Languidness ([PILL] Pennebaker, 1982). On
the PILL participants were asked to report the frequency with which they had experienced 54
common physical symptoms and sensations (e.g., ‘eyes water’) on a five-point scale ranging
from 0 (Have never or almost never experienced the symptom) to 4 (More than once every
week). Higher scores indicate greater levels of bodily sensation reporting.
2.2.3 Interoception control tasks
The Water Quantity Estimation test was designed to match the WLT and required
participants to estimate the amount of water contained in three identical, but unusually
contoured bottles (See Fig. 2), at three separate time periods similar to those used in the WLT.
Each 150 ml bottle contained a different amount of water: Time one: 75mls; Time two: 37.5
mls; and Time three: 100 mls. Without touching the bottles, participants were asked to
estimate the absolute fullness level in the bottles on a 15cm VA scale (anchors: ‘Empty’ to
‘Full’). At Times Two and Three, participants were also asked to estimate the volume change
in the bottle, relative to the previous bottle on a 15cm VA scale (anchors: ‘A lot less’ to ‘A lot
more’). Data for test here were the relative and absolute volume values generated by
participants.

The second interoceptive control task was the Tone Counting task, which was matched to
the HRT task methodology. This involved the participants silently counting a regular
metronome tone over four counterbalanced time periods of 25, 35, 45, and 55 seconds. After
initially identifying a level at which the tone could be comfortably heard, participants
completed a 20-second practice session in order to familiarise with the task. The tone was set

11

at 72 beats per minute, and after completing each time interval, the participants reported the
number of tones heard to the researcher. Each counting interval was separated by a 30 second
resting period during which the participants were instructed to sit quietly without conversing.
The metronome tone was computer generated (TempoPerfect Metronome v 3.08 NCH
Software) and participants listened via headphones. Finally, a single absolute deviation value
was calculated for the Tone Counting task in the same manner as for the HRT.
2.2.4 Neuropsychological tests
Logical Memory (LM), a subtest of the Wechsler Memory Scale – Fourth Edition
(WMS-IV; Wechsler, 2009), is a measure of immediate and delayed memory for structured
verbal information (two stories) and is sensitive to MTL damage (e.g., Griffith et al., 2004). In
order to minimise the well-documented practice effects, which can occur with repeated
exposure to LM (e.g., Schnabel 2012), an alternate form (Sullivan, 2005) was administered.
This was particularly important for JW and PT, who had had recent prior exposure to the
Wechsler LM stories. Both alternate stories had similar characteristics to the WMS-IV LM
passages (e.g., number of words, characters, sentences; Sullivan, 2012) and were administered
using the WMS-IV standardised instructions.
The Digit Span subtest (Forward, Backwards, and Sequencing) of the Wechsler Adult
Intelligence Scale – Fourth Edition (WAIS-IV; Wechsler, 2008) was included as a measure of
attention, concentration and working memory to ensure that these constructs were intact in all
participants. This test was administered as per the manual and the Digit Span Total Scaled
Score (SS) is reported.
The Test of Pre-Morbid Functioning (TOPF) was used to estimate patient and control IQ
(Wechsler, 2011) and was administered and scored as per standardised instructions.
2.3 Procedure

12

All participants were asked to abstain from eating and drinking for two hours prior to
testing. All reported complying with this request. After completing the information and
consent form, the test battery was administered. Test ordering for the study was fixed (see
Table 1) as some tests had to precede others as they had multiple parts.

2.4 Analysis
Data were analysed using Modified Crawford’s t-tests (Crawford & Howell, 1998), with
alpha set at .05. For the key test of whether any impairment on the HRT task is less than any
impairment on the WLT fullness ratings, scores on the HRT task and the WLT (linear and
quadratic coefficients for fullness ratings) were all converted to Z-scores (relative to controls).
The difference in performance between the standardised HRT task and the standardised linear
(and quadratic) coefficient fullness scores for WLT were calculated for each patient, and for
the controls, to test for any differential deficit across test type.

3. Results
3.1 Biographical and background questionnaire data
There were no significant differences between JW and PT and controls in age, education,
BMI, DASS scores, exercise (hours/week; IPAQ), or sleep quality (PIRS) – see Table 2.

3.2 Neuropsychological testing
As expected, both JW and PT scored significantly below the control mean in the LM
delayed recall (see Table 3). PT also showed significant impairment in the immediate recall

13

condition. On Digit Span (see Table 3), there were no significant differences between JW and
PT, and the controls. There were also no differences in estimated premorbid IQ (see Table 3).

3.3 Interoception tasks
No difference was expected between patients and controls on the HRT task. However,
while JW performed somewhat less accurately than controls, PT’s HRT score was very poor
and significantly worse than controls (see Table 4). No participant reported using any
calculation based strategy to generate a response rather all reported trying to focus on their
heart rate. We note that across the whole sample (including both patients and excluding the
one control who provided no responses on this task) the greater the number of trials on which a
response was provided the more accurate the heart rate estimation (r = 0.49, p < 0.01).
We predicted that patient reports of fullness would differ from controls on the WLT,
noting that we found no difference in water intakes (see Table 4). For the linear coefficient
data, PT’s fullness ratings decreased significantly more over time than controls (see Fig. 3a),
but JW’s did not differ from controls (see Fig. 3b). Turning to the quadratic coefficients, JW’s
score was significantly different from controls, with far more dynamic changes in fullness
across time (See Fig. 3b). No significant difference was seen between PT’s quadratic
coefficient and that of controls (see Fig. 3a).

The key test is whether any impairment on the HRT task is less than any impairment on
the WLT fullness ratings. Crucially, patients’ degree of impairment did not differ between the
two interoception tasks: for JW, HRT task vs. WLT linear coefficient, t = 1.63, HRT task vs.
WLT quadratic coefficient, t = 0.6; for PT, HRT task vs. WLT linear coefficient, t = 0.18, HRT

14

task vs. WLT quadratic coefficient, t = 1.41. Thus, contrary to predictions, there was no
evidence of any differential deficit in performance between the HRT task and the WLT fullness
ratings in our two patients.
As there was no differential deficit between the HRT task and the WLT (fullness ratings),
two overall deficit scores were computed. One was the average of the HRT task Z score and
the WLT linear coefficient Z score, and the other was the average of the HRT task Z score and
the WLT quadratic coefficient Z score. These were calculated for each patient, with these
scores then compared to scores calculated for the controls.
For JW, there was a significant overall impairment for the joint HRT/WLT quadratic
coefficient score, t(27) = 2.49, p < 0.01, but not for the HRT/WLT linear coefficient score (t <
1). PT was abnormal on both comparisons, for the HRT/WLT linear and quadratic scores,
respectively, t(27) = 4.64, p < 0.01 and t(27) = 2.37, p < 0.01.
Finally, we tested whether there was any difference on the questionnaire measure of bodily
sensations (PILL). No significant differences in symptom reporting were found between JW
and PT, relative to controls (see Table 4).
3.3 Interoception control tasks
No significant differences were found between patients and controls on the ability to detect
the absolute amounts of fluid contained in the bottles over the three time periods (JW all t’s <
0.6; PT all t’s < 1.78). There were also no differences between patients and controls in their
capacity to estimate the relative volume change between the bottles at T1 vs T2 and T2 vs T3
(JW both t’s < 0.32; PT both t’s < 0.96).
There were no significant differences between JW and PT, and the control group on the
number of tones counted during each time period (t’s < 0.88).

4. Discussion

15

The current study investigated whether two individuals with selective damage to MTL
structures, but an intact IC and ACC, would demonstrate differential impairments across two
types of interoceptive task. The WLT was hypothesised to involve the integration of current
physiological state (i.e., gastric distension) with mnemonic information (i.e., fluid consumed).
We predicted that performance on this interoception task would be impaired in individuals with
MTL lesions. In contrast, the HRT task was hypothesised not to require integration, and prior
fMRI research suggested that the IC and ACC would mediate performance on this task, not
MTL structures. Thus, no impairment was expected in the MTL patients here. Overall, the
two patients, JW and PT were poor at estimating their heart-rate, with PT’s result being
significantly worse than controls. As expected, both PT and JW demonstrated abnormal
patterns in their fullness ratings over time, even though their water intake did not differ
significantly from the control group’s. Crucially, there was no evidence of a differential deficit
between JW and PTs performance on the WLT relative to the HRT task. Rather, there was a
general impairment across both tasks - relative to controls. These results suggest that task
differences are unlikely to explain the divergent findings that emerge between lesion studies in
patients and animals on the one hand, and neuroimaging in healthy subjects on the other.
Looking more specifically at the patients’ interoceptive data, we note that their general
pattern of performance was abnormal. Both JW and especially PT were inaccurate when it
came to tracking their own heart rates. On the WLT, both PT and JW reported significantly
different fullness responses relative to controls, although the differences identified for each
patient were not the same. PT’s ratings started high (i.e., full) and then decreased across the
study – even after drinking (see Fig. 2). In contrast, JW’s ratings started low (i.e., empty) and
increased very rapidly after drinking (from almost the bottom to almost the top of the scale)
after drinking only a modest amount (150ml) of water. These differences in pattern between
PT and JW are not readily explained unless both are just reflective of guessing due to some

16

degree of interoceptive agnosia. When compared to HM (Hebben et al., 1985) and the other
MTL cases (Higgs et al., 2008; Rozin et al., 1998), where ratings of hunger and fullness were
obtained, these were characterised by an unresponsiveness to food intake (i.e., hunger and
fullness ratings remained static). Some evidence of this can also be seen in PT, but even her
ratings decrease at the final time point. Such differences could be attributed to the degree of
MTL damage. This is far more limited in our two patients, relative to the severe form (with
dense anterograde amnesia) in the other studied patients (Hebben et al., 1985; Higgs et al.,
2008; Rozin et al., 1998). Finally, while on the topic of lesion extent, it is noteworthy that the
unilateral lesion in PT appeared to be as effective as the bilateral lesion in JW at generating
interoceptive impairments. Interestingly, Craig (2009) suggests a right-sided dominance
(albeit for insula activation) for mediating particular types of interoception that are arousing,
including heart beat awareness and gastric distension.
A number of other tests were also conducted. JW and PT’s reports of attentiveness to a
range of internal bodily sensations (e.g., indigestion, constipation, nausea) did not differ
significantly to that of controls, which was perhaps surprising given their performance on the
WLT and HRT task. One possibility is that responses to the PILL could be based on the
patients’ experiences prior to MTL damage, rather than on current interoceptive ability as
presumably measured by the WLT and the HRT task. No differences were observed between
patients and controls on either of the two interoceptive control tasks, suggesting that general
task-related deficits (e.g., counting or volume estimation) were unlikely to explain the
interoception findings. In regards to memory - and as would be expected - no differences were
observed in working memory performance, but both JW and PT’s immediate and delayed
verbal memory scores were poorer than those of the controls.
The essential idea motivating this research was that the brain areas identified as being
responsible for interoception may differ according to the type of task. Studies using patients or

17

animals with brain lesions have often used tasks that involve integrating current feeling states
with mnemonic processes, while neuroimaging research with healthy people has often used
tasks that do not. As both JW and PT were, overall, impaired on both types of interoceptive
task (relative to the controls), this suggests that task type may not be a promising explanation
for the contrasting findings that come from lesion and neuroimaging studies. Below, we
briefly consider three further possible explanations of this difference.
A first and obvious starting point is to reconsider our assumption that the HRT task does
not involve a task-specific mnemonic component. As we noted in the introduction, it is
possible that under certain circumstances participants may feel forced to rely upon some form
of explicit calculation, utilising memory of their past heart rate (e.g., Ring et al., 2015). We
attempted to forestall this possibility by modifying the HRT task, so that when participants felt
utterly unable to generate a response there was no pressure to provide one. Using this
procedure we found that failures to report a tracking response were significantly associated
with poorer cardiac interoceptive accuracy on reported trials, suggesting that those unable to
consciously perceive their heart beating were indeed those with the lowest score on this task.
In addition, no participant indicated using any form of recollection/calculation approach, not
surprisingly when there was no pressure to do so.
A related possibility is that participants draw upon some form of implicit memory of
cardiac output to inform their judgment (see Azevedo et al., 2016). For example, estimating
ones heartbeat may be difficult as interoception may be intermittent, and so people may rely
upon some form of implicit memory of their heart’s rhythm so that they can keep counting
over periods when they cannot consciously perceive their heartbeat. While traditionally it was
thought that MTL structures played little role in implicit memory processes, more recent data
suggests they do, and so it is plausible that an intact MTL may be necessary for this taskspecific process to occur (Hannula & Greene, 2012). Importantly, however, this rests upon

18

two untested assumptions: (1) that participants do use some form of implicit memory to
generate an estimate of their own heart rate on the HRT task; and (2) that the task-specific
implicit memory process is MTL dependent. Notwithstanding these caveats, we return later to
the more general idea that the brain may form models of interoceptive state that are memory
dependent.
A second explanation for the divergent outcomes between much of the interoceptionrelated fMRI data and the lesion data (human and animal) derives from a network view of
interceptive processing. Several authors (e.g., Phillips et al., 2003a & b; Schoenbaum, et al.,
2002) have suggested that the limbic system is composed of two parts: The ventral system
(amygdala, anterior IC, caudal orbital cortex and ventral parts of the ACC), which identifies
the emotional significance of a stimulus and the resultant affective state, and the dorsal system
(dorsal cingulate gyrus, hippocampus and PFC), which regulates these states. This division
could suggest why MTL lesions may impair interoception in the absence of generally finding
MTL activity. While these two systems are separate, it is plausible that together they form part
of a multidimensional map of current interoceptive state, which then guides goal-directed
behaviors. It may be that the MTL, and especially the hippocampus, reflects a unique point of
vulnerability for this whole network, even if MTL structures are not always active.
A third possibility concerns the way in which much of the interoception neuroimaging
research has been conducted. One pertinent example of this concerns the only fMRI study of
the HRT task (Pollatos et al., 2007b), which found no MTL activations – in contrast to the
conclusion from this study. One reason for this type of result may be the insensitivity of fMRI
to temporal changes between the functionally connected brain networks active during
interoceptive tasks (Jarrahi et al., 2015). Using deconstructed fMRI data and multivariate
approaches Jarrahi et al., (2015) studied the temporal activation of neural networks in response
to bladder stimulation. Bladder fullness engaged multiple structures, including the IC, ACC,

19

amygdalo-hippocampi and parahippocampal gyri, with interactions between the different areas
changing across time. Indeed, when this type of technique is applied to cardiac interoception, a
similar picture emerges, and one that is far more consistent with the data from this and other
lesion studies. Kuehn et al., (2016) grouped participants into accurate or inaccurate cardiac
interoceptors based upon the HRT task. Participants were then imaged and asked to just focus
on their heart (i.e., not to count or do anything else). Here significant alterations were observed
in functional connectivity between the insula and hippocampus, which differed between
accurate and inaccurate cardiac interoceptors. It may be then, that it is the type of imaging
paradigm that is responsible for the difference in findings/conclusions between fMRI and
lesion studies, rather than the type of interoceptive task.
Finally, it is important to consider what functional role MTL structures play in
interoception. This manuscript was initially motivated by the idea that MTL involvement
might be restricted to certain tasks that required participants to integrate a memory of prior
events (e.g., amount of food consumed or fluid drunk) with judgments of current state. We
suggested that the Water Loading Task involved integration while the cardiac interoception
task did not. The finding here of impaired cardiac interoception in MTL patients argues
against this type of account. A second possibility is that MTL structures and especially the
hippocampus, record interoceptive experience across time in a context-dependent manner,
which then forms a crucial part of the input into the type of comparator model of interoception
advanced by Barrett and Simmons (2015). In the comparator model, the brain constantly
predicts interoceptive state, based upon past state information, and then compares this with
current input. Feedback about the success or failure of the current model is then returned to the
inputting system(s). Barrett and Simmons (2015) model utilises structures identified by
existing interoceptive models (i.e., Craig, 2009), notably the anterior cingulate cortex and IC,
but does not include roles for other structures that have been identified from lesion and

20

functional connectivity studies (i.e., MTL). Including MTL structures as an input of past state
information would be an important extension of the comparator model. In particular, it would
allow context to modulate input to the model, so that previous interoceptive experience at a
particular place, time, mood etc was used to make current state predictions. In addition, we
would suggest that a comparator model could not function properly in the absence of such
context dependent input, which may be why MTL damage results in such significant
impairment to interoception.
In conclusion, this study demonstrates that selective damage to MTL structures - with
intact IC and ACC - can impair interoception, irrespective of the type of task. These findings
challenge the view derived from neuroimaging literature, which suggests that the IC and ACC
mediate interoceptive processing. Rather, we suggest, the MTL or an MTL centred network, is
intimately involved in supporting interoception possibly via its involvement in some form of
comparator model.

References
Annese, J., Schenker-Ahmed, N. M., Bartsch, H., Maechler, P., Sheh, C., Thomas, N., ...
Klaming, R. (2014). Postmortem examination of patient HM’s brain based on
histological sectioning and digital 3D reconstruction. Nature Communications, 5.doi:
10.1038/ncomms4122
Azevedo, R., Aglioti, S. & Lenggenhager, B. (2016). Participants’ above chance recognition of
own-heart sound combined with poor metacognitive awareness suggests implicit
knowledge of own cardiodynamics. Scientific Reports, 6, 26545.
Barrett, L. & Simmons, W. (2015). Interoceptive predictions in the brain. Nature Reviews
Neurosciences, 16, 419-429.

21

Berman, S. M., Naliboff, B. D., Suyenobu, B., Labus, J. S., Stains, J., Ohning, G., ... Mayer, E.
A. (2008). Reduced brainstem inhibition during anticipated pelvic visceral pain correlates
with enhanced brain response to the visceral stimulus in women with irritable bowel
syndrome. The Journal of Neuroscience, 28, 349-359. doi: 10.1523/JNEUROSCI.250007.2008
Brunstrom, J. M., Burn, J. F., Sell, N. R., Collingwood, J. M., Rogers, P. J., Wilkinson, L. L.,
... Ferriday, D. (2012). Episodic memory and appetite regulation in humans. PloS One, 7,
e50707. doi: 10.1371/journal.pone.0050707
Canavan, A. G. M., Passingham, R. E., Marsden, C. D., Quinn, N., Wyke, M., & Polkey, C. E.
(1989). Sequencing ability in Parkinsonians, patients with frontal lobe lesions and
patients who have undergone unilateral temporal lobectomies. Neuropsychologia, 27,
787-798. doi: 10.1016/0028-3932(89)90003-1
Clifton, P. G., Vickers, S. P., & Somerville, E. M. (1998). Little and often: ingestive behavior
patterns following hippocampal lesions in rats. Behavioral Neuroscience, 112, 502-511.
doi: 10.1037/0735-7044.112.3.502
Craig, A. D. (2002). How do you feel? Interoception: the sense of the physiological condition
of the body. Nature Reviews Neuroscience, 3, 655-666. doi: 10.10.38/nrn894
Craig, A. D. (2003). Interoception: the sense of the physiological condition of the body.
Current Opinion in Neurobiology, 13, 500-505. doi: 10.1016/S0959-4388(03)00090-4
Craig, A. D. (2009). How do you feel—now? The anterior insula and human awareness.
Nature Reviews Neuroscience, 10, 59-70. doi: 10.1038/nrn2555
Craig, C. L., Marshall, A. L., Sjöstrom, M., Bauman, A, E., Booth, M. L., Ainsworth, B. E., . . .
Oja, P. (2003). International physical activity questionnaire: 12-country reliability and
validity. Medicine & Science in Sports & Exercise, 195, 3508-1381. doi:
10.1249/01.mss.0000078924.61453.fb.

22

Crawford, J. R., & Howell, D. C. (1998). Comparing an individual's test score against norms
derived from small samples. The Clinical Neuropsychologist, 12, 482-486. doi:
10.1076/clin.12.4.482.7241
Davidson, T. L., Chan, K., Jarrard, L. E., Kanoski, S. E., Clegg, D. J., & Benoit, S. C. (2009).
Contributions of the hippocampus and medial prefrontal cortex to energy and body
weight regulation. Hippocampus, 19, 235. doi: 10.1002/hipo.20499
Davidson, T.L., and Jarrard, L.E. (1993). A role for hippocampus in the utilization of hunger
signals. Behavioral and Neural Biology, 59, 167-171. doi: 10.1016/0163-1047(93)909258
Delparigi, A., Chen, K., Salbe, A., Hill, J., Wing, R., Reiman, E. & Tataranni, P. (2004).
Persistence of abnormal neural responses to a meal in post-obese individuals.
International Journal of Obesity, 28, 370-377.
Denton, D., Shade, R., Zamarippa, F., Egan, G., Blair-West, J., McKinley, M., ... Fox, P.
(1999). Neuroimaging of genesis and satiation of thirst and an interoceptor-driven theory
of origins of primary consciousness. Proceedings of the National Academy of Sciences,
96(9), 5304-5309.
Duddu, V., Isaac, M. K., & Chaturvedi, S. K. (2003). Alexithymia in somatoform and
depressive disorders. Journal of Psychosomatic Research, 54, 435-438. doi:
10.1016/S0022-3999(02)00440-3
Geeraerts, B., Van Oudenhove, L., Dupont, P., Vanderghinste, D., Bormans, G., Van Laere, K.
& Tack, J. (2011). Different regional brain activity during physiological gastric
distension compared to balloon distension: A H2 O-PET study. Neurogastroenterology
and Motility, 23, 533-545.

23

Greer, S. M., Goldstein, A. N., & Walker, M. P. (2013). The impact of sleep deprivation on
food desire in the human brain. Nature Communications, 4, 1-7. doi:
10.1038/ncomms3259
Griffith, H. R., Pyzalski, R. W., Seidenberg, M., & Hermann, B. P. (2004). Memory
relationships between MRI volumes and resting PET metabolism of medial temporal
lobe structures. Epilepsy & Behavior, 5, 669-676. doi: 10.1016/j.yebeh.2004.06.020
Hannula, D. & Greene, A. (2012). The hippocampus reevaluated in unconscious learning and
memory: at a tipping point? Frontiers in Human Neuroscience, 6, Article 80.
Hebben, N., Corkin, S., Eichenbaum, H., & Shedlack, K. (1985). Diminished ability to
interpret and report internal states after bilateral medial temporal resection: case HM.
Behavioral Neuroscience, 9, 1031-1039.
Herbert, B. M., Muth, E. R., Pollatos, O., & Herbert, C. (2012). Interoception across
modalities: on the relationship between cardiac awareness and the sensitivity for gastric
functions. PloS One, 7, e36646. doi: 10.1371/journal.pone.0036646
Herbert, B. M., & Pollatos, O. (2014). Attenuated interoceptive sensitivity in overweight and
obese individuals. Eating Behaviors, 15, 445-448. doi: 10.1016/j.eatbeh.2014.06.002
Herbert, B. M., Pollatos, O., Flor, H., Enck, P., & Schandry, R. (2010). Cardiac awareness and
autonomic cardiac reactivity during emotional picture viewing and mental stress.
Psychophysiology, 47, 342-354. doi: 10.1111/j.1469-8986.2009.00931.x
Herbert, B. M., Pollatos, O., & Schandry, R. (2007). Interoceptive sensitivity and emotion
processing: an EEG study. International Journal of Psychophysiology, 65, 214-227. doi:
10.1016/j.ijpsycho.2007.04.007
Higgs, S., Williamson, A. C., Rotshtein, P., & Humphreys, G. W. (2008). Sensory-specific
satiety is intact in amnesics who eat multiple meals. Psychological Science, 19, 623-628.
doi: 10.1111/j.1467-9280.2008.02132.x

24

Jarrahi, B., Mantini, D., Balsters, J. H., Michels, L., Kessler, T. M., Mehnert, U., & Kollias, S.
S. (2015). Differential functional brain network connectivity during visceral
interoception as revealed by independent component analysis of fMRI time‐ series.
Human Brain Mapping, 36, 4438-4468. doi: 10.1002/hbrn.22929
Jeneson, A., & Squire, L. R. (2012). Working memory, long-term memory, and medial
temporal lobe function. Learning & Memory, 19, 15-25. doi: 10.1101/lm.024018.111
Komisaruk, B. R., Whipple, B., Crawford, A., Grimes, S., Liu, W. C., Kalnin, A., & Mosier, K.
(2004). Brain activation during vaginocervical self-stimulation and orgasm in women
with complete spinal cord injury: fMRI evidence of mediation by the vagus nerves. Brain
Research, 1024, 77-88. doi: 10.1016/j.brainres.2004.07.029
Kuehn, E., Mueller, K., Lohmann, G. & Schuetz-Bosbach, S. (2016). Interoceptive awareness
changes the posterior insula functional connectivity profile. Brain Structure and
Function, 221, 1555-1571.
Lovibond, P. F., & Lovibond, S. H. (1995). The structure of negative emotional states:
Comparison of the Depression Anxiety Stress Scales (DASS) with the Beck Depression
and Anxiety Inventories. Behaviour Research and Therapy, 33, 335-343. doi:
10.1016/0005-7967(94)00075-U
Mehnert, U., Boy, S., Svensson, J., Michels, L., Reitz, A., Candia, V., ... Schurch, B. (2008).
Brain activation in response to bladder filling and simultaneous stimulation of the dorsal
clitoral nerve—An fMRI study in healthy women. Neuroimage, 41, 682-689. doi:
10.1016/j.neuroimage.2008.03.006
Michal, M., Reuchlein, B., Adler, J., Reiner, I., Beutel, M., Vogele, C., Schachinger, H. &
Schulz, A. (2014). Striking discrepancy of anomalous body experiences with normal
interoceptive accuracy in depersonalization-derealization disorder. PLoS One, 9, e89823.

25

Moul, D.E., Pilkonis, P.A., Miewald, J.M., Carey, T.J. & Buysse, D.J. (2002). Preliminary
study of the test-retest reliability and concurrent validities of the Pittsburgh Insomnia
Rating Scale (PIRS). Sleep, 25, A246-A247.
Pennebaker, J. W. (1982). The psychology of physical symptoms. New York: Springer-Verlag.
Phillips, M. L., Drevets, W. C., Rauch, S. L., & Lane, R. (2003). Neurobiology of emotion
perception I: The neural basis of normal emotion perception. Biological Psychiatry, 54,
504-514. doi: 10.1016/S0006-3223(03)00168-9
Phillips, M. L., Gregory, L. J., Cullen, S., Cohen, S., Ng, V., Andrew, C., ... Thompson, D. G.
(2003). The effect of negative emotional context on neural and behavioural responses to
oesophageal stimulation. Brain, 126, 669-684. doi: 10.1093/brain/awg065
Pollatos, O., Gramann, K., & Schandry, R. (2007a). Neural systems connecting interoceptive
awareness and feelings. Human Brain Mapping, 28, 9-18. doi: 10.1002/hbm.20258
Pollatos, O., Schandry, R., Auer, D. P., & Kaufmann, C. (2007b). Brain structures mediating
cardiovascular arousal and interoceptive awareness. Brain Research, 1141, 178-187. doi:
10.1016/j.brainres.2007.01.026
Pollatos, O., Traut‐ Mattausch, E., & Schandry, R. (2009). Differential effects of anxiety and
depression on interoceptive accuracy. Depression and Anxiety, 26, 167-173. doi:
10.1002/da.20504
Ring, C., Brener, J., Knapp, K. & Mailloux, J. (2015). Effects of heartbeat feedback on beliefs
about heart rate and heartbeat counting: A cautionary tale about interoceptive awareness.
Biological Psychology, 104, 193-198.
Rosenkranz, M. A., Busse, W. W., Johnstone, T., Swenson, C. A., Crisafi, G. M., Jackson, M.
M., ... Davidson, R. J. (2005). Neural circuitry underlying the interaction between
emotion and asthma symptom exacerbation. Proceedings of the National Academy of

26

Sciences of the United States of America, 102, 13319-13324. doi:
10.1073/pnas.0504365102
Rozin, P., Dow, S., Moscovitch, M., & Rajaram, S. (1998). What causes humans to begin and
end a meal? A role for memory for what has been eaten, as evidenced by a study of
multiple meal eating in amnesic patients. Psychological Science, 9(5), 392-396.
Schandry, R. (1981). Heart beat perception and emotional experience. Psychophysiology,
18(4), 483-488.
Schnabel, R. (2012). Overcoming the challenge of re-assessing logical memory. The Clinical
Neuropsychologist, 26, 102-115. doi: 10.1080/13854046.2011.640642
Schoenbaum, G., Setlow, B., Saddoris, M. P., & Gallagher, M. (2003). Encoding predicted
outcome and acquired value in orbitofrontal cortex during cue sampling depends upon
input from basolateral amygdala. Neuron, 39, 855-867. doi: 10.1016/S08966273(03)00474-4
Schulz, A., Lass-Hennemann, Sutterlin, S., Schachinger, H. & Vogele, C. (2013). Cold pressor
stress induces opposite effects on cardioreceptive accuracy dependent on assessment
paradigm. Biological Psychology, 93, 167-174.
Scoville, W. B., & Milner, B. (1957). Loss of recent memory after bilateral hippocampal
lesions. Journal of Neurology, Neurosurgery, and Psychiatry, 20, 11-21. doi:
10.1136/jnnp.20.1.11
Squire, L., Stark, C. & Clark, R. (2004). The medial temporal lobe. Annual Review of
Neuroscience, 27, 279-306.
Sullivan, K. (2005). Alternate forms of prose passages for the assessment of auditory–verbal
memory. Archives of Clinical Neuropsychology, 20, 745-753. doi:
10.1016/j.acn.2005.04.006

27

Vaitl, D. (1996). Interoception. Biological Psychology, 42, 1-27. doi: 10.1016/03010511(95)05144-9
Van Oudenhove, L., Demyttenaere, K., Tack, J., & Aziz, Q. (2004). Central nervous system
involvement in functional gastrointestinal disorders. Best Practice & Research Clinical
Gastroenterology, 18, 663-680. doi: 10.1016/j. bpg. 2004.04.010
Wang, G. J., Tomasi, D., Backus, W., Wang, R., Telang, F., Geliebter, A., ... Volkow, N. D.
(2008). Gastric distention activates satiety circuitry in the human brain. Neuroimage, 39,
1824-1831. doi: 10.1016/j.neuroimage.2007.11.008.
Wechsler, D. (1987). Wechsler Memory Scale-Revised. New York: The Psychological
Corporation.
Wechsler, D. (1997). Wechsler Memory Scale-Third Edition. Pearson; San Antonio, TX.
Wechsler, D. (2008). Wechsler Adult Intelligence Scale–Fourth Edition. Pearson; San Antonio,
TX.
Wechsler, D. (2009). Wechsler Memory Scale–Fourth Edition. Pearson; San Antonio, TX.
Wechsler, D. (2011). Test of Premorbid Functioning. Pearson; San Antonio, TX.

Fig.1: MRI Coronal T2 Flair images of (a) Case PT and (b) Case JW. From left to right,
images move more posteriorly.
Fig.2: Bottles used in the Water Quantity Estimation Task
Fig.3: (a) Fullness ratings for PT and controls (mean) from T1-T3 and (b) Fullness ratings for
JW and controls (mean) from T1-T3

Fig.1
(a)

28

(b)

29

Fig.2

30

Fig.3a & b

Fullness ratings 0-15

15

10
Controls
5

PT

0
1

2

3

Time Periods

Fullness ratings 0-15

15

10
Controls

5

JW
0
1

2

3

Time Periods

Table 1: Order of task administration

Order

Task

1

Biographical and Background Questionnaire data

31

2

Water Load Test (WLT) - Hunger, Thirst, Fullness Ratings #1

3

Water Quantity Estimation test #1

4

Heart Rate Tracking (HRT) Task #1

5

Logical Memory Immediate Recall

6

WLT – Drinking phase

7

WLT - Hunger, Thirst, Fullness Ratings #2

8

Test of Premorbid Functioning

9

Water Quantity Estimation test #2

10

Depression, Anxiety, Stress Scale–21 (DASS-21)

11

Digit Span (Forward, Backwards, Sequencing)

12

Logical Memory Delayed Recall

13

HRT Task repeat #2

14

Pennebaker Inventory of Limbic Languidness (PILL)

15

Water Quantity Estimation test #3

16

Tone Counting task

17

WLT - Hunger, Thirst, Fullness Ratings #3

Table 2: Biographical and background questionnaire data for MTL patients and
controls, and comparison of patients to controls

Variable

Age

Patient data

Control data

Crawford t-test

JW

PT

M (SD)

JW

PT

61

39

47 (10.9)

1.30

0.74

32

BMI

18.5

29.8

25.3 (5.2)

1.34

0.87

Education

3

3

4.8 (0.8)

0.95

0.95

Depression

6

2

4.5 (5.4)

0.28

0.48

Anxiety

0

0

4.5 (5.0)

0.92

0.92

Stress

4

2

9.4 (7.3)

0.76

1.04

5.3

11.3 (15.2)

0.39

0.40

2

2.1 (1.4)

1.60

0.10

DASS-21

Exercise (IPAQ)
5.5
Sleep (PIRS) 0

Table 3: Neuropsychological data for MTL patients and controls, and
comparison of patients to controls

Variable

Patient data

Control data

Crawford t-test

JW

PT

M (SD)

JW

PT

2.89*

Logical Memory
Immediate

5

3

9.2 (2.2)

1.95

Delayed

5

2

10.3 (2.4)

2.06* 2.90*

11

11.6 (2.9)

0.14

0.21

94

109.7 (10.3)

1.42

1.55

Digit span (total)
Scaled score 12
TOPF
IQ estimate 124

33

* p < 0.05
Table 4: Interoceptive task data for MTL patients and controls, and comparison
of patients to controls

Task
Variable

Patient data

Control data

Crawford t-test

JW

M (SD)

JW

PT

0.67 (0.21)

1.29

2.44*

-0.2 (2.4)

1.32

2.51*

PT

Heart rate tracking (HRT) task
HRT score 0.42

0.08

Water loading task (WLT)
Linear coefficient (fullness ratings)
3.0

-6.3

Quadratic coefficient (fullness ratings)
-10.4 -5.8

-3.3 (3.1)

2.31* 0.82

464.0 (202.6)

1.58

0.57

0.87

0.03

Water intake (ml)
150

350

Pennebaker Inventory of Limbic Languidness (PILL)
PILL Score 20

12

13.2 (7.9)

* p < 0.05

Highlights


fMRI and lesion studies suggest different brain regions support interoception



We tested whether this arises from the type of interoceptive task used

34


Medial temporal lobe (MTL) lesion patients were compared to controls on two tasks



Patients were equally impaired on both tasks, irrespective of its need for memory



The role of the MTL in interoception and the contrast with fMRI data are discussed