J Neurol Neurosurg Psychiatry 1998;65:771–773 771

J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.65.5.771 on 1 November 1998. Downloaded from http://jnnp.bmj.com/ on 26 March 2019 by guest. Protected by copyright.
SHORT REPORT

Involvement of the human cerebellum during

habituation of the acoustic startle response:
A PET study
Dagmar Timmann, Christina Musso, Florian P Kolb, Michel Rijntjes, Markus Jüptner,
Stefan P Müller, Hans Christoph Diener, Cornelius Weiller

Abstract the human cerebellum in habituation of the

The present study investigated the in- acoustic startle response was investigated in
volvement of the human cerebellum in the healthy subjects using PET.
habituation of the acoustic startle re-
sponse using PET. The startle response Materials and methods
was elicited in seven young, healthy sub- With the approval of the local ethics committee
jects by a tone presented via headphones. the study was performed in seven healthy sub-
Startle responses were recorded from the jects (five men and two women; age from 25 to
right sternocleidomastoid muscle. Re- 63 years, mean 35.5 years) without any hearing
gional cerebral blood flow (rCBF) was loss, without neurological and orthopaedic dis-
assessed in nine scans and one startle eases, and not receiving any medication. All
stimulus was applied during each scan. subjects were right handed. All gave their
The reduction of size of the sternocleido- informed consent.
Department of mastoid muscle response was correlated
Neurology with changes in rCBF during the ongoing STARTLE RESPONSE HABITUATION
D Timmann process of startle response habituation. A
M Jüptner Subjects were lying supine and with their eyes
H C Diener
significant decrease of rCBF was found in closed. Head position was maintained by use of
the medial cerebellum. These data are an individually molded foam headrest to mini-
Department of consistent with an involvement of the mise involuntary head movements during the
Nuclear Medicine, medial parts of the human cerebellum in scans. The startle stimulus consisted of an
University of Essen, non-associative learning as proposed by
Hufelandstraâe 55,
auditory tone pulse (1000 Hz; 95 dB; 50 ms)
previous animal studies. applied bilaterally via headphones. This tone
45122 Essen, Germany (J Neurol Neurosurg Psychiatry 1998;65:771–773)
S P Müller was superimposed on continuous white noise
Keywords: cerebellum; habituation; startle response;
(50 dB SPL) to mask environmental noise.
Institute of Physiology, Electromyography recordings were per-
positron emission tomography
University of Munich, formed from the right sternocleidomastoid
Pettenkoferstraâe 12,
80336 Munich,
muscle with surface electrodes (sampling
Germany The role of the cerebellum for associative and rate=1 kHz). The EMG traces were amplified,
F P Kolb non-associative learning has been intensively full wave rectified and bandpass filtered (10
studied in animals. Classical conditioning of Hz<f<10 kHz). Latency and integrated EMG
Department of the nictitating membrane reflex has often been activity (iEMG) were calculated after visual
Neurology, identification of the onset of the sternocleido-
Friedrich-Schiller-
used to study the role of the cerebellum for
University of Jena, associative learning and habituation of the mastoid muscle activity using a commercial
Philosophenweg 3, acoustic startle response for non-associative analysis program (axograph®). Due to diYcul-
07740 Jena, Germany learning.1 2 Deficits in classical conditioning of ties in defining the end point of the sternoclei-
C Musso the eyeblink and limb flexion reflex have been domastoid muscle burst the EMG integral was
M Rijntjes demonstrated in patients with cerebellar disor- calculated using a fixed interval of 100 ms after
C Weiller the onset of the EMG burst. Sternocleidomas-
ders and in healthy subjects using PET.3 4
Correspondence to: In normal subjects, unexpected auditory toid muscle iEMG was normalised by assign-
Dr Dagmar Timmann, stimuli produce a startle response which ing an arbitrary value of 100% to each subject’s
Department of Neurology, usually involves flexion of the neck and trunk iEMG in the first trial and referencing changes
University of Essen,
Hufelandstraâe 55, 45122 and brief closure of the eyes.5 The activity in their iEMG to that value. The ongoing pro-
Essen, Germany. Telephone responsible for the startle response originates cess of habituation was defined as the trial by
0049 201 723 2460; fax 009 within the lower brain stem.5 6 Animal studies trial reduction in size of the sternocleidomas-
201 723 5901; email toid muscle response.
Dagmar.Timmann@
have shown that the medial cerebellum is part
uni-essen.de of the essential circuitry for long term habitua-
tion of the acoustic startle response.6–8 So far, SCANNING METHODS
Received 15 December 1997 the role of the cerebellum in habituation of the Regional cerebral blood flow (rCBF) was
and in revised form
20 April 1998 startle response has not been investigated in measured during nine consecutive scans using
Accepted 23 April 1998 humans. In the present study, involvement of the H215O injection technique with an ECAT
772 Timmann, Musso, Kolb, et al

J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.65.5.771 on 1 November 1998. Downloaded from http://jnnp.bmj.com/ on 26 March 2019 by guest. Protected by copyright.
t Maps showing all cerebellar areas correlating with the process of acoustic startle reflex habituation (p<0.05). The average
PET data (n=7) were superimposed on to the MRI template. The highest correlation was seen in the intermediate parts of
the left cerebellar hemisphere (p=0.014).

953–15 PET scanner (CTI Inc, Knoxville, TN, cess of startle response habituation (p<0.05,
USA) over 90 seconds. Ten seconds after the corrected for multiple comparisons). The
beginning of each scan one startle stimulus was cerebellar area of significantly decreasing rCBF
applied (= total of nine auditory tone bursts). extended from the vermis to intermediate
Data acquisition and analysis have been regions of the cerebellar hemispheres predomi-
described previously in detail.4 In brief, assess- nately on the left side (figure). The highest cor-
ment of significant rCBF change was per- relation was seen in the left intermediate part of
formed using statistical parametric mapping the cerebellum (maximum: x=−20 mm, y=−44
(SPM, Wellcome Department of Cognitive mm, z=−32 mm; Z=4.44; p=0.014 (corrected
Neurology, Queen Square, London, UK). The for multiple comparisons)). Local maxima in
scans were transformed into the standard the area of significant change of regional
stereotactic anatomical space of Talairach and cerebellar blood flow correlating with startle
Tournoux.9 Because of the restricted field of response habituation are given in the table.
view of our camera (5.4 cm) we decided to scan There were no brain areas outside the cerebel-
the lower parts of the brain which included the lum which showed significant change of rCBF
cerebellum. Scans were performed from −56 correlating with the process of habituation.
mm below the AC-PC line for cerebellar planes However, it has to be emphasised that the view
to +24 mm above the AC-PC line for the fron- of our camera was restricted (5.4 cm) and only
tal lobes, as the camera was tilted. the lower parts of the brain were scanned.
The rCBF of each test scan (1–9) was corre-
lated pixel by pixel with normalised sternoclei-
domastoid muscle iEMG of each individual Discussion
subject elicited by the startle stimulus during We showed significant decreases in rCBF
the corresponding scan using the general linear correlating with the process of habituation of
model.10 Significant changes in rCBF were the human acoustic startle response in the
assumed at p<0.05 (corrected for multiple medial parts of the cerebellum. These data are
comparisons). Results are displayed as statisti- consistent with an involvement of the human
cal parametric maps showing the significance cerebellum in the habituation of the startle
level of areas of significant change of rCBF. response.
The cerebellar vermis has been shown to be
a key region of a circuit essential for the acqui-
Results sition of long term habituation of the acoustic
Subjects habituated the acoustic startle re- startle response in rats.6–8 However, medial cere-
sponse over the course of nine startle stimuli. bellar damage did not disrupt the basic neural
Mean sternocleidomastoid muscle activity was circuitry for the acoustic startle response itself
reduced to 50.1% (SD=19.0) of its initial size or the short term habituation mechanism
by the end of the PET session. Decrease of which is presumed to be intrinsic to that path-
sternocleidomastoid muscle iEMG was way within the caudal brainstem.11 Because the
roughly linear (linear regression: R=0.70, cerebellum is not involved in the basic neural
slope=−3.5 , p<0.05). circuit for the acoustic startle response, in-
The correlation analysis showed an area of volvement in the modulation of the response
significantly decreased blood flow in the size seems likely.12 The significant decrease of
cerebellum correlating with the ongoing pro- cerebellar activity during the process of habitu-
Local maxima in the area of significant decrease (p<0.05) of regional cerebellar blood flow ation might reflect a decrease in tonic activity
correlating with startle response habituation on reticular neurons subserving startle re-
sponses during the process of habituation.13 In
Cerebellar location x y z Z value the present study, eVects of short term and long
Intermediate cerebellum (left) −20 −44 −32 4.44* term habituation were not separated with the
Vermis (right) 8 −54 −24 3.44 subjects receiving a total of nine startle stimuli
Vermis (left) −10 −70 −28 3.37
Intermediate cerebellum (left) −20 −60 −36 3.17
within one 2 hour test session. The significant
Intermediate cerebellum (right) 24 −42 −26 3.03 change of rCBF might reflect activity related to
Intermediate cerebellum (left) −36 −46 −22 2.71 short term or long term habituation.
Intermediate cerebellum (left) −40 −60 −28 2.52
It has to be emphasised that PET studies
*p=0.014. alone cannot tell us whether a particular locus
Cerebellum and startle response 773

J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.65.5.771 on 1 November 1998. Downloaded from http://jnnp.bmj.com/ on 26 March 2019 by guest. Protected by copyright.
is essentially involved in a function or informed non-associative learning as suggested by previ-
about a function located elsewhere. The ous animal studies. Further studies are needed
presented data of activity changes may most to clarify if the human cerebellum is involved in
simply reflect the monitoring of peripheral short term, or long term habituation,or both,
events, one of the cerebellar tasks generally.14 and whether its involvement is essential or not.
As the size of muscular responses is diminished
during the process of habituation the reduced This paper was supported in part by a DFG-grant (Dagmar
activity in the cerebellum might simply reflect Timmann, Ti 239/2–1), by an EU grant (CT94/1261), and a
grant from the Kuratorium ZNS Bonn. We gratefully thank
the reduced motor output. H-G Elles for technical assistance, B Terschüren for her help in
The present study provides further evidence conducting the experiments, S Kiebel for his help in analysing
the data, and M Weiâ for preparing the manuscript.
for the idea of functional compartments within
the human cerebellum. A previous study from
1 Lavond DG, Kim JJ, Thompson RF. Mammalian brain sub-
this laboratory showed a significant change of strates of aversive classical conditioning. Annual Review of
activity of the ipsilateral intermediate part of Psychology 1993;44:317–42.
2 Bloedel JR, Bracha V. On the cerebellum, cutaneomuscular
the human cerebellum during classical condi- reflexess, movement control and the elusive engrams of
tioning of the human flexion reflex.4 During memory [review]. Behav Brain Res 1995;68:1–44.
3 Woodruf-Pak D. Classical conditioning. International Review
startle response habituation, the area of signifi- of Neurobiology 1997;41:341–66
cantly decreased rCBF included the vermis 4 Timmann D, Kolb FP, Baier C, et al. Cerebellar activation
during classical conditioning of the human flexion reflex: a
and adjacent intermediate parts of the cerebel- PET-study. Neuroreport 1996;7:2056–60.
lum. The vermis seems predominantly to be 5 Brown P, Rothwell JC, Thompson PD, et al. New
observations on the normal auditory startle reflex in man.
involved in non-associative learning (habitua- Brain 1991;114:1891–902.
tion of the startle response) whereas diVerent 6 Leaton RN, Supple WF. Cerebellar vermis: essential for
long-term habituation of the acoustic startle response. Sci-
parts of the intermediate cerebellum contrib- ence 1986;232:513–5.
ute to non-associative and associative learning 7 Leaton RN, Supple WF. Medial cerebellum and long-term
habituation of acoustic startle in rats. Behav Neurosci 1991;
(classical conditioning). These findings are 105:804–16.
supported by animal studies which showed dif- 8 Lopiano L, De’Sperati C, Montarolo PG. Long-term
habituation of the acoustic startle response: role of the cere-
ferent localisation for specific aversive classical bellar vermis. J Neurosci 1990;35:79–84.
conditioning and non-specific fear habituation 9 Talairach J, Tournoux P. Co-planar stereotaxic atlas of the
human brain. New York: Thieme, 1988.
or conditioning: Cerebellar vermal lesions pre- 10 Friston KJ, Holmes AP, Worsley KJ, et al. Statistical
vented long term startle reflex habituation and parametric maps in functional imaging: a general approach.
Human Brain Mapping 1995;2:189–210.
heart rate conditioning but not eyelid condi- 11 Leaton RN, Cassella JV, Borszcz GS. Short term and long
tioning, whereas lateral cerebellar lesions term habituation of acoustic startle response in chronic
decerebrate rats. Behav Neurosci 1985;99:901–12.
prevented eyelid conditioning but not heart 12 MacKay WA, Murphy JT. Cerebellar modulation of reflex
rate conditioning.1 6–8 15 gain. Prog Neurobiol 1979;13:361–417.
13 Eccles JC, Nicoll RA, Rantucci T, et al. Topographic studies
In summary, involvement of the medial cer- on medial reticular nucleus. J Neurophysiol 1976;39:109–
ebellum in the habituation of the startle 18.
14 Kolb FP, Rubia FJ, Bauswein E. Comparative analysis of
response has been demonstrated in human cerebellar unit discharge patterns in the decerebrate cat
subjects by changes in rCBF, monitored via during passive movements. Exp Brain Res 1987;68:219–33.
15 Supple WF, Leaton RN. Lesions of the cerebellar vermis
PET. These data are consistent with an and cerebellar hemispheres: eVect on heart rate condition-
involvement of the human cerebellum in ing in rats. Behav Neurosci 1990:104:934–47.