Brain (1996), 119, 1249-1253
Bilateral simultaneous transcranial Doppler
monitoring of flow velocity changes during
visuospatial and verbal working memory tasks
Letizia M. Cupini,1 Maria Matteis,2 Elio Troisi,2 Maurizio Sabbadini,2 Giorgio Bernardi,1'2
Carlo Caltagirone,1-2 and Mauro Silvestrini1
1
Clinic of Neurology, S. Eugenio Hospital-University of
Rome 'Tor Vergata' and the ~IRCCS 'S. Lucia', Rome, Italy
Correspondence to: Dr L. M. Cupini, Clinica Neurologica
Universita di Roma 'Tor Vergata', Ospedale S. Eugenio,
P.le dell'Umanesimo 10, 00144 Roma, Italy
Summary
Twenty-two normal volunteers were studied by simultaneous
bilateral transcranial Doppler (TCD) monitoring of flow
velocity changes in the middle cerebral arteries during
visuospatial and verbal working memory tasks. All tasks
were visually presented material; they required full
engagement of subjects during the study, similar attention
demands and a manual response when the appropriate target
was recognized. Stimuli were presented tachistoscopically.
Each task produced significant and distinct effects on flow
velocity with respect to baseline values. When side-to-side
comparisons were performed, a significantly higher increase
in flow velocity was observed in the right middle cerebral
artery (MCA) compared with the left one during the
visuospatial task (P < 0.005). Our findings agree with results
of recent studies using PET which show lateralized activation
during working memory tasks. The high temporal resolution
of this technique suggests that further application in cerebral
flow change monitoring during neuropsychological studies
could be promising.
Keywords: transcranial Doppler; working memory
Abbreviations: MCA = middle cerebral artery; MFV = mean flow velocity; TCD = transcranial Doppler
Introduction
Many clinical and experimental data suggest that memory
cannot be viewed as an unitary function: a distinction
between short-term and long-term processes is no longer
disputed (Atkinson and Shiffrin,
1968). Cognitive
neuroscientists suggest that within short-term memory
(working memory) different subsystems can be. recognized
(Desimone, 1992). Baddeley (1992) suggests that we possess
a working memory system for storing and manipulating new
information for brief periods of time and proposes that
working memory consists of a central executive system
co-ordinating information from two slave systems: (i) the
visuospatial sketch pad, which manipulates visual images;
(ii) the phonological loop, which stores and rehearses speechbased information. The brain areas essentially involved in
the working memory system are not well known.
With TCD ultrasonography it is possible to detect
instantaneous changes in the flow velocity of large cerebral
arteries. Even if this information cannot be used for describing
© Oxford University Press 1996
absolute CBF values, it can be considered a reliable indicator
of rapid flow changes and then of perfusion in the territory
supplied by the large intracerebral arteries (Aaslid et al.,
1982; Bishop et al., 1986; Aaslid, 1987; Kontos, 1989).
The aim of the present study was to investigate the effects
of both visual and verbal working memory tasks on CBF
changes by means of simultaneous bilateral monitoring of
mean flow velocity (MFV) using the TCD technique.
Subjects and methods
After giving informed consent, 22 right-handed volunteers
(13 men, nine women; mean age: 25.6 years, range 19-32
years) took part in the study. Right-handedness was assessed
by the Edinburgh handedness inventory (Oldfield, 1971).
No subjects had a history of neurological disorder, their
literacy level was high and Italian was their first language.
The study was carried out in a quiet darkened room with
1250
L. M. Cupini et al.
(A)
Table 1 Verbal working memory task and control task
(1) Phonological short-term memory task
Randomized sequences of six phonologically dissimilar
consonants of the Italian alphabet were presented on a screen
at the rate of one per second. Subjects were instructed to
rehearse the stimuli silently and to remember them. Two
seconds after each sequence, a probe consonant appeared and
subjects judged whether it was present in the previous
sequence. Subjects responded by pushing the button once
with their left hand to indicate whether the probe was present
in the previous sequence. They made no response if the
probe was not previously present.
500 ms
3.000 ms
200 ms
1.500 ms
Fig. 1 (A) Visuospatial working memory task (memory
condition). Subjects were instructed to fixate a cross in the centre
of the screen for 500 ms, the cross was then supplemented for
200 ms by three dots appearing on the circumference of an
imaginary circle centred on the cross, followed by a 3000 ms
period during which a cross appeared. Finally, a probe encircled
the location of one of the previous dots or not for 1500 ms.
Subjects responded by pushing the button once with their right
hand to indicate whether the probe marked the location of a dot.
They made no response if the probe did not mark the location of
a dot. (B) Control task (perception condition). Subjects fixated a
cross in the centre of the screen for 3500 ms. The cross was then
supplemented for 200 ms by three dots appearing on the
circumference of an imaginary circle centred on the cross,
followed by a 1500 ms period during which three dots and the
probe circle were presented simultaneously. Subjects responded
by pushing the button once with their right hand to indicate
whether the probe encircled a dot. They made no response if the
probe did not encircle a dot. Adapted from fig. 1 in Jonides et al.
(1993).
subjects sitting in a comfortable position. Subjects were
asked to refrain from speaking and moving, to relax and
breathe regularly during both rest phases and mental tasks.
Mean flow velocity in the right and left MCA was
continuously and simultaneously monitored by means of a
Multi-Dop X/TCD 7 TCD instrument (DWL Elektronische
Systeme GmbH, Germany; Esaote Biomedica, Italy). Two
dual 2-MHz transducers, fitted on a headband and placed on
the temporal bone window, were used to obtain a bilateral
continuous measurement. The highest signal was sought at a
depth ranging from 46 to 54 mm. This unit allowed
continuous-wave Doppler recording of the intracranial arteries
with on-line calculation of MFV in cm s~'. By activating the
record function it is possible to save the Doppler spectra for
the entire period of each study. In all subjects, recording was
performed during a 1-min rest phase in which subjects were
instructed to empty their mind and during the performance
of each mental task.
Mental tasks were taken from earlier PET studies (Paulesu
et al., 1993; Jonides et al., 1993) and included (i) a
visuospatial working memory task (Fig. 1); (ii) a verbal
working memory task (Table 1). In addition, the protocol
included a control condition for each task. They were
respectively a perception condition (Fig. 1) and a control
task, which could not be transcoded phonologically (Table
(2) Control task
This task was identical to the verbal task with the exception
that Korean letters (not phonologically codeable) were used.
1). Each subject performed the tasks sequentially. After each
task, before proceeding to the recording of a new rest and
mental task condition, a return to baseline values of flow
velocity was documented. Tasks were explained via a tape
recording played before the study. Stimuli were projected
onto a screen mounted on the wall. The stimuli were shown
by means of a slide projector with a fixed speed shutter
mounted on the lens to allow for tachistoscopic presentation.
The speed of stimuli presentation is shown in Table 1 and
Fig. 1. Subjects responded manually at the end of each item
using a push button to indicate recognition of the appropriate
targets. They were instructed to work as quickly and
accurately as possible. An observer monitored their response
accuracy. The order of administration of conditions was
counterbalanced across subjects. All subjects responded with
the hand homolateral to the hemisphere presumably involved
in the task. The right hand was utilized during the perception
condition and the visuospatial tasks, and the left hand during
the verbal working memory task. In this way, we intended
to prevent the coexisting effect of contralateral MFV increase
during motor activity (button pushes) (Silvestrini et al., 1993).
In order to check the possible effects of changing pCO2
on vessel diameter, end-tidal COj was monitored during the
study by means of a CO2 analyser (Normocap-oxy, Datex,
Italy). Mean blood pressure reading and heart rate were
continuously monitored by means of a blood pressure monitor
(2300 Finapress Ohmeda, Col., USA). Recordings of mean
blood pressure and heart rate (averaged over a 60-s period)
were made at the end of each phase (rest and mental
task) of the study. Recorded values of these two variables
represented the mean of mean blood pressure and heart rate
values during the different phases of the study, i.e. the rest
phase and the mental task.
Data for analysis consisted of the mean of the recorded
values of MFV, mean blood pressure and heart rate during
the rest phases and mental tasks. Absolute values of the
changes were considered.
Analysis of MFV changes for each mental task was
performed by means of a two-way ANOVA (dependent
factor: MFV changes) with side of recording (right versus left
Flow velocity changes during cognitive activity
Table 2 Baseline and activation values for MFV in the
right and left MCAs during the performance of mental
tasks
Mental task
Italian letter
Rest
Task
Korean letter
Rest
Task
Visuospatial memory
Rest
Task
Visuospatial perception
Rest
Task
Right MCA
Left MCA
65.3 (8.7)
71.7(9.3)
64.8 (8.3)
71.3 (9.7)
66.3(11.2)
71.5 (10.8)
65.6 (10.9)
70.6(11.5)
63.9(12.1)
74.0(13.0)
64.1 (10.0)
68.2 (10.3)
65.8 (13.4)
70.2 (12.6)
66.2 (10.0)
69.2(9.1)
Mean flow velocities are expressed in cm s '. Standard deviations
are reported in parentheses.
MCA) and condition (rest versus mental task performance) as
within factors. The two-way ANOVA was repeated
introducing mean blood pressure and heart rate absolute
changes as covariant factors to evaluate the possible effects
of these two variables on blood flow velocity changes.
Moreover, to compare MFV changes during memory tasks
and their control conditions, a three-way ANOVA was
performed with factors: (i) side of recording; (ii) memory
(memory task versus control task); (iii) modality (visuospatial
versus verbal). Finally, since different hands were used for
the different tasks, we performed a further analysis to rule
out the possible hand effect on MFV changes. We subtracted
MFV changes of each control condition from the homologous
memory task in which, as already stated, the same hand
(right hand for the visuospatial task and left hand for the
verbal task) had been used. These values were then used to
compare visuospatial and verbal task by means of a two-way
ANOVA with side (right versus left MCA) and differences in
MFV changes calculated as specified above as within factors.
The statistical package used was the Complete Statistical
System with graphics and data management (Statsoft Inc.,
Tulsa, Okla., USA, 1988).
Results
Mean values of MFV at rest and during task performance
are reported in Table 2. All tasks were associated with a
significant bilateral increase in MFV compared with the rest
phase. The interaction sideXcondition was significant only
during the visuospatial working memory task [/•"(!,42) =
8.6; P < 0.005]. This interaction was due to the fact that
the task produced a significantly higher increase in MFV
compared with the baseline value in the right than in the left
MCA (Table 2). This was also confirmed by the three-way
ANOVA that showed that side of recording (right versus
left), memory (memory task versus control task) and modality
(visuospatial versus verbal task) interaction was significant
1251
[F(l,42) = 5.2; P < 0.03]. A post hoc comparison (Tukey
test) showed that this was due to the fact that only during
the visuospatial memory task was the increase of MFV
significantly higher in the right MCA than in the left MCA.
On the contrary, no side-to-side differences were observed
during the other tasks.
The use of different hands in different tasks did not
influence the results since, when the differences of MFV
changes between modality (visuospatial and verbal tasks)
and their control conditions were considered to compare the
two memory tasks (two-way ANOVA), a right-side activation
still gave a significant result [F(l,42) = 5.2; P < 0.03]
during the visuospatial task.
There are obvious similarities in the CBF velocity pattern
of changes observed during the visuospatial location task
compared with the Korean letter task. However, only during
the former task were statistically significant difference
between right and left side observed.
Subjects were quite accurate in their behavioural
performance during letter recognition and during the visual
memory task; they made no errors in the perception condition.
Both mean blood pressure and heart rate showed slight
similar increases during each mental task.
End-tidal pCO2 did not show significant modification,
with values oscillating from 38 to 42 mmHg throughout
each recording. When the two-way ANOVA (side of
recordingXcondition) was repeated, introducing mean
blood pressure, heart rate changes and pCO2 as covariated
factors, the results of analysis remained substantially
unchanged.
Discussion
The main purpose of this study was to confirm and extend
earlier observations (Markus and Boland, 1992; Silvestrini
et al., 1994; Rihs et al., 1995) concerning the high sensitivity
of TCD in detecting different patterns of MFV changes
during the performance of distinct mental tasks. We attempted
to study CBF velocity changes induced by working memory
tasks. The term working memory refers to a brain system
that provides temporary storage and manipulation of the
information necessary for such complex cognitive tasks as
language comprehension, learning and reasoning. It is distinct
from our ability to store and recall the innumerable facts and
events that accumulate over a lifetime (declarative memory)
and it is thought to be composed of a central executive
system coordinating information from two slave systems, the
visuospatial sketch pad and the phonological loop (Baddeley,
1992). Recently, Paulesu et al. (1993) and Jonides et al.
(1993) used PET to measure CBF changes during the
performance of working memory tasks and provided maps
of the brain area involved. In their study, Jonides et al.
(1993) observed that a visuospatial memory task activates
right-sided structures, including right inferolateral, prefrontal
and lateral premotor cortices, as well as right inferior parietal
and dorsolateral occipital cortices. In Paulesu's study, verbal
1252
L. M. Cupini et al.
working memory tasks engaged the two components of the
'articulatory loop', the subvocal rehearsal system associated
with activation of the inferoposterior frontal region (Broca's
area), and the phonological store, associated with activation
of the left inferior parietal cortex.
We observed a distinct pattern of activation associated
with each kind of task. During the recovery phase, a fast
return to baseline values was always observed. As expected,
visuospatial memory task induced a significant increase of
MFV on the right side. This finding agrees with the results
of Jonides et al. (1993). Letter verbal task induced a significant
bilateral increase of MFV compared with baseline values,
but no significant differences between the left and right side.
This finding is at variance with the results from previous
studies in man (Paulesu et al., 1993) in which a significant
activation of the left hemisphere was observed with PET
analysis during a letter verbal task. However, other
authors (Raife et al., 1992; Grasby et al, 1993; Petrides
et al., 1993) have observed a bilateral activation when
measuring rCBF during the performance of verbal working
memory tasks. Even in the Paulesu et al. (1993) study, a
nonsignificant increase in homologous structures was
recognized, implying a contribution of these regions to
task performance. The significative bilateral activation we
observed is likely to reflect the visual analysis of the written
stimuli. Different problem solving strategies (phoneme-tographeme transcoding process) may account for the
variability in activation patterns observed across subjects. In
fact, the analysis of the single pattern of activation revealed
the presence of two different groups of subjects: those with
strictly unilateral activation (right or left side) and those
with bilateral activation (with only a slight nonsignificant
prevalence for right or left side). Finally, in both perception
condition and Korean letter task, a slight nonsignificant
higher increase of MFV was observed in the right MCA with
respect to the left one. This finding suggests the visual
involvement of such tasks. In the present study, we utilized
tasks very similar to those used in the PET studies of Paulesu
et al. (1993) and Jonides et al. (1993) to reproduce a similar
experimental condition. All tasks were visually presented
material and were similar in terms of rate of stimulus
presentation. All tasks required full engagement of subjects
during the study. The control tasks were designed not only
to reproduce the finger movements used in the test condition
but also to engage in attention demanding similar to that
required during the experimental tasks. We were also
interested in obtaining information regarding full engagement
of subjects in studies and their accuracy during the
performance of each task. As is well known, vocalization
can produce artifacts during TCD monitoring (Diehl et al.,
1990). For this reason, we utilized tasks requiring manual
responses when the appropriate targets were recognized.
The main limit of TCD application is related to the low
spatial resolution. However, the characteristics of TCD, in
particular lack of invasiveness, low cost, good tolerability,
high sensitivity in detecting instantaneous changes in CBF
velocity and the possibility of simultaneous recordings of
MFV of both hemispheres, obtainable with the newgeneration machines, suggest further utilization of this
technique in CBF velocity change monitoring during
neuropsychological studies. In the present study, by coupling
TCD recording and tachistoscopic presentation of stimuli,
we were able to confirm the possibility of correlating cerebral
artery flow dynamics with selective cerebral hemispheric
activation during the execution of tasks requiring the activity
of the right or left hemisphere. Based on these considerations,
TCD cannot be proposed as an alternative to other techniques
of CBF investigation, such as PET and single photon emission
tomography. By means of the latter approaches, it is possible
to investigate flow changes at regional intrahemispheric levels
and so to obtain information regarding cerebral localization
of distinct functions. On the other hand, TCD ensures the
possibility of following rapid flow modification in real time
in response to different stimulation. This characteristic seems
to be more useful in the study of correlations among reaction
time, performance, accuracy in response to distinct stimuli
and CBF changes. In this respect, this technique has been
preliminary utilized in the evaluation of the recovery from
motor and aphasic problems in patients with cerebral
ischaemic lesions (Silvestrini et al., 1993, 1995). The
peculiarity of the physiopathological information supported,
outlines the possibility of a wider employment of TCD in
different physiological and pathological situations involving
the CNS.
References
Aaslid R. Visually evoked dynamic blood flow response of the
human cerebral circulation. Stroke 1987; 18: 771-5.
Aaslid R, Markwalder TH, Nornes H. Noninvasive transcranial
Doppler ultrasound recording of flow velocity in basal cerebral
arteries. J Neurosurg 1982; 57: 769-74.
Atkinson RC, Shiffrin RM. Human memory: a proposed system
and its control processes. In: Spence KW, Taylor Spence J, editors.
The psychology of learning and motivation. Advances in research
and theory, Vol. 2. New York: Academic Press, 1968: 89-195.
Baddeley A. Working nemory. [Review], Science 1992; 255: 556-9.
Bishop CC, Powell S, Rutt D, Browse NL. Transcranial Doppler
measurement of middle cerebral artery blood flow velocity: a
validation study. Stroke 1986; 17: 913-15.
Desimone R. The physiology of memory: recordings of things past.
Science 1992; 258: 245-6.
Diehl RR, Sitzer M, Hennerici M. Changes of cerebral blood flow
velocity during cognitive activity [letter; comment]. Stroke 1990;
21: 1236-7. Comment on: Stroke 1989; 20: 1005-11.
Grasby PM, Frith CD, Friston KJ, Bench C, Frackowiak RSJ, Dolan
RJ. Functional mapping of brain areas implicated in auditory-verbal
memory function. Brain 1993; 116: 1-20.
Jonides J, Smith EE, Koeppe RA, Awh E, Minoshima S, Mintun
MA. Spatial working memory in humans as revealed by PET [see
Flow velocity changes during cognitive activity
1253
comments]. Nature 1993; 363: 623-5. Comment in: Nature 1993;
363: 583-4.
study of verbal working memory [abstract]. Neurosci Abstr Soc
1992; 18: 932.
Kontos HA. Validity of cerebral arterial blood flow calculations
from velocity measurements. Stroke 1989; 20: 1-3.
Rihs F, Gutbrod K, Gutbrod B, Steiger H-J, Sturzenegger M, Mattle
HP. Determination of cognitive hemispheric dominance by 'stereo'
transcranial Doppler sonography. Stroke 1995; 26: 70-3.
Markus HS, Boland M. 'Cognitive activity' monitored by noninvasive measurement of cerebral blood flow velocity and its
application to the investigation of cerebral dominance. Cortex 1992;
28: 575-81.
Silvestrini M, Caltagirone C, Cupini LM, Matteis M, Troisi E,
Bernardi G. Activation of healthy hemisphere in poststroke recovery:
a transcranial Doppler study. Stroke 1993; 24: 1673-7.
Oldfield RC. The assessment and analysis of handedness: the
Edinburgh inventory. Neuropsychologia 1971; 9: 97-113.
Silvestrini M, Cupini LM, Matteis M, Troisi E, Caltagirone C.
Bilateral simultaneous assessment of cerebral flow velocity during
mental activity. J Cereb Blood Flow Metab 1994; 14: 643-8.
Paulesu E, Frith CD, Frackowiak RSJ. The neural correlates of the
verbal component of working memory [see comments]. Nature
1993; 362: 342-5. Comment in: Nature 1993; 363: 583-4.
Petrides M, Alivisatos B, Meyer E, Evans AC. Functional activation
of the human frontal cortex during the performance of verbal
working memory tasks. Proc Natl Acad Sci USA 1993; 90: 878-82.
Raife EA, Fiez JA, Raichle ME, Balota DA, Petersen SE. A PET
Silvestrini M, Troisi E, Matteis M, Cupini LM, Caltagirone C.
Involvement of the healthy hemisphere in recovery from aphasia
and motor deficit in patients with cortical ischemic infarction. A
transcranial Doppler study. Neurology 1995; 45: 1815-20.
Received January 5, 1996. Revised March 6, 1996.
Accepted March 23, 1996