Hemodynamic monitoring of middle cerebral arteries during cognitive tasks performance

Authors

  • Marina Boban MD, PhD,

    Corresponding author
    1. Department of Neurology, University Hospital Centre, Zagreb, Croatia
    2. School of Medicine, University of Zagreb, Zagreb, Croatia
    • Correspondence: Marina Boban, MD, PhD, University Department of Cognitive Neurology, Zagreb School of Medicine and University Hospital Centre, Kišpatićeva 12, HR-10000 Zagreb, Croatia. Email: maboban@mef.hr

    Search for more papers by this author
  • Petra Črnac MD,

    1. School of Medicine, University of Zagreb, Zagreb, Croatia
    Search for more papers by this author
  • Anamari Junaković,

    1. Department of Neurology, University Hospital Centre, Zagreb, Croatia
    Search for more papers by this author
  • Branko Malojčić MD, PhD

    1. Department of Neurology, University Hospital Centre, Zagreb, Croatia
    2. School of Medicine, University of Zagreb, Zagreb, Croatia
    Search for more papers by this author

Abstract

Aim

The aim of this study was to obtain temporal pattern and hemispheric dominance of blood flow velocity (BFV) changes and to assess suitability of different cognitive tasks for monitoring of BFV changes in the middle cerebral arteries (MCA).

Methods

BFV were recorded simultaneously in MCA during performance of phonemic verbal fluency test (pVFT), Trail Making Tests A and B (TMTA and TMTB) and Stroop tests in 14 healthy, right-handed volunteers aged 20–26 years.

Results

A significant increase of BFV in both MCA was obtained during performance of all cognitive tasks. Statistically significant lateralization was found during performance of Stroop test with incongruent stimuli, while TMTB was found to have the best activation potential for MCA.

Conclusion

Our findings specify TMTB as the most suitable cognitive test for monitoring of BFV in MCA.

An increase of regional cerebral blood flow (rCBF) during performance of cognitive tasks was described in the published work a long time ago.[1] Excellent correlation of relative rCBF and relative blood flow velocity (BFV) changes enables good insight into regional metabolic demands at the microcirculatory level by measuring BFV in large cerebral arteries.[2] Transcranial Doppler (TCD) ultrasonography provides non-invasive monitoring of BFV in intracranial large arteries on different anatomical levels.[2-4] High temporal resolution, safety and accessibility are some of the advantages of TCD in comparison to other functional imaging methods, such as functional magnetic resonance imaging (fMRI), single-photon emission computed tomography, and positron emission tomography (PET).[5] On the other hand, one of the major disadvantages is low spatial resolution determined by the size of the cortical area supplied by the investigated artery.[5]

Monitoring of BFV changes in large intracranial arteries during performance of different cognitive tasks has been widely investigated.[6-9] During performance of a verbal fluency test (VFT), a bilateral, statistically significant increase of BFV in the middle cerebral arteries (MCA) was found with significant left lateralization,[9-12] confirming results of functional neuroimaging studies that have shown activation of the left dorsolateral prefrontal cortex, premotor cortex, and anterior cingulated cortex during performance of VFT.[13, 14] So far, no other tests used in this study (Trail Making Tests [TMT] and Stroop tests) have been investigated in TCD studies.

TMT, as tests of visual attention and task switching, can provide information about scanning, visuomotor tracking, mental flexibility and speed of processing and therefore activate wide areas of the brain with emphasis on the frontal and temporal cortex, as well as the cingulated gyrus.[15, 16] Another group of tests used in this study, Stroop tests, provide information about selective attention, mental flexibility and processing speed. Frontal lobe activation, particularly in the left anterior cingulate and dorsolateral prefrontal cortex and bilateral activation of superior and inferior parietal lobulus have been found during performance of these tests.[17, 18]

The aim of our study was to obtain temporal patterns of BFV changes and to determine hemispheric dominance in MCA during pVFT, TMTA, TMTB and Stroop tests performance. Additionally, we aimed to assess the most suitable cognitive task for monitoring of BFV changes in MCA.

Methods

Subjects

The study was designed as an exploratory study and was conducted at the Department of Neurology, University Hospital Centre Zagreb and was approved by the local ethics committee. Informed consent was obtained from all subjects participating in the study. Fourteen native Croatian right-handed healthy volunteers aged 20–26 years with normal or adequately corrected visual acuity were included in the study. Abstinence from caffeine and nicotine 12 h before testing was another inclusion criterion. Exclusion criteria were left-handedness and history of neurological, psychiatric or cardiac disease.

Methods

All participants completed a general questionnaire (with questions regarding their age, education, current and previous diseases) and the Edinburgh Handedness Inventory (EHI) questionnaire (adapted from Oldfield[19]). BFV in both MCA and heart rate (HR) were monitored during performance of cognitive tasks presented on a computer screen. Instructions for completing the paradigm were described beforehand in detail to each subject verbally and were displayed on the computer screen before each task. Blood pressure, HR and visual analog anxiety scale (VAS) were assessed before and after completing the paradigm. During completion of the paradigm, participants were sitting in front of a computer screen in a semi-dark, quiet room.

Experimental paradigm

The paradigm consisted of five cognitive tasks and lasted approximately 65 min. Between each task there was a resting period of 2 min that proved (on preliminary testing) to be sufficient for BFV and HR to return to basal levels. During resting periods, the participants were looking at a conventional screensaver on the computer screen (Starfield, Microsoft, Redmond, WA, USA).

The initial phase was resting period with the subject sitting relaxed and watching the computer screen for 180 s. It was followed by a breath-holding test (BHT) lasting for up to 30 s with a subsequent 120-s resting period when subjects were asked to breathe normally. BHT was performed as a screening test for identifying subjects with impaired vasomotor reactivity. The breath-holding index was calculated (BHI = (BFVend – BFVref)/t; BFVend being the BFV at the end of breath-holding, BFVref being the BFV velocity during the −15 to −3-s prestimulus period of the resting interval,[20] and t being the duration of breath holding in s). BHI between 1.03 and 1.65 was another inclusion criteria for the study.[21]

  1. pVFT. Participants were instructed to pronounce silently as many words as they could recollect in 25 s beginning with the given letter (with the exception of proper names). Five seconds later they were asked to say out loud the number of generated words.[11] The task consisted of 10 subtasks (with one rehearsal subtask), differing by the given letter, with a 1-min resting period between subtasks.
  2. Stroop test with neutral stimulus. This consisted of silent reading of color names written in black letters on the computer screen within 15 s. It included four subtasks (with a 1-min resting period between subtasks) differing by the order of color names.
  3. Stroop test with congruent stimulus. Groups of six identically colored ‘X’ symbols were displayed differing in displayed color; subjects were instructed to silently nominate the color of each ‘XXXXXX’ group within 20 s. It included four subtasks (with a 1-min resting phase between them) with differing color orders.
  4. Stroop test with incongruent stimulus. A list of color names was displayed on the screen with mismatch of the color name and font color. The participant was instructed to nominate the font color disregarding the written color name over 25 s paying attention to accuracy and speed. It was comprised of a trial task and eight subtasks divided by 1-min resting intervals.
  5. TMT. Two subtasks (TMTA and TMTB) were performed.[22] The TMTA consisted of circles numbered 1–25 distributed over a sheet of paper. Subjects were instructed to connect the numbers in ascending order by drawing lines on the paper within 15 s. TMTB included both numbers (1–13) and letters (A–L). Subjects were instructed to connect as many numbers and letters by lines in ascending sequence over 45 s with alternating numbers and letters (1–A–2–B–3–C–4–D …–13.) Each subtask was preceded by a shorter trail task. Subtasks were divided by 2-min resting interval and performed on a piece of paper.

TCD monitoring

A 2-MHz pulse-wave TCD device (Doppler-BoxX, Compumedics DWL, Singen, Germany) was used. The probes were fixed to the headframe bilaterally over a transtemporal window at a 50–65-mm insonation depth related to M1 segments of the MCA.[3, 23] Temporal resolution of the recording was 0.01 s and the data were recorded as a tab delimited text. Mean BFV (MBFV) was calculated by the TCD device using the following equation: MBFV = (Vsis + 2Vdis)/3; Vsis being peak systolic velocity and Vdis peak diastolic velocity.

Averaging of MBFV values for each task and each subject was made for 5-s intervals with 0.5 s overlapping. Finally averaged MBFV of 5-s lasting periods were obtained for each cognitive task for the whole group of subjects. HR was calculated by counting the distance between peak systolic velocities on a diagram during the same time intervals. Due to the vast difference in actual values of MBFV among participants, relative MBFV (MBFVrel) was calculated by the following equation: MBFVrel = (MBFVact/MBFVref) × 100; MBFVact being the actual MBFV during the observed interval, and MBFVref being the MBFV velocity during the −15 to −3-s prestimulus period of the resting interval.[20]

Statistics

MBFV values were analyzed and averaged (described in the section TCD monitoring, Methods) for each MCA and activation (BHT and cognitive) task. Data are presented as line charts showing relative MBFV changes during different 5-s time intervals for BHT and each cognitive task. The Kolmogorov–Smirnov test (with Lilliefors correction) was used to confirm that clinical parameters (age, education, EHI index, VAS score and blood pressure values before and after monitoring) as well as MBFV values were normally distributed. Data are presented as mean ± SD.

Significance of MBFV increase during the performance of each cognitive task was assessed by using t-test for dependent samples.

Laterality indices (LI) were assessed by defining the 5-s time interval within activation period with both the highest and most significant MBFV difference between left and right MCA. Afterwards, three 5-s intervals before and three 5-s intervals after this interval were picked and compared (seven time interval points for each MCA) by using t-test for matched pairs. Left MCA dominance was defined by positive LI, and right MCA dominance was defined by negative LI.

T-test for matched pairs was also used for comparison of activation potential of each cognitive task.

Spearman's test was used to assess the correlation between MBFV values and HR by measuring correlation coefficient (r). P-values below 0.05 were considered statistically significant. The analysis was performed by using spss 11.0.1 (spss, Chicago, IL, USA).

Results

Subjects

Fourteen right-handed, highly educated subjects (seven female and seven male) were included in the study (age 23.33 ± 1.23 years; education 16.26 ± 1.10 years). Regarding VAS, systolic and diastolic pressure, results revealed no statistically significant differences before and after task performance.

MBFV

Resting period: there was no statistically significant difference in MBFV between right MCA (R-MCA) and left MCA (L-MCA) during the 2-min resting phase.

BHT: a rapid statistically significant increase of MBFV in both MCA was recorded within the first 5 s (a 10.98 ± 8.17% increase in R-MCA and 14.37 ± 7.35% increase in L-MCA compared to the reference period) followed by decrease of MBFV reaching minimal levels in the time period from 10th to 15th s (10–15 s; a drop of 16.64 ± 8.05 in R-MCA and of 17.18 ± 8.82 in L-MCA). Subsequent slow increase of MBFV followed reaching a maximum in the time period 35–40 s after task start (a 30.28 ± 9.05% increase in the R-MCA and 25.19 ± 8.25% increase in the L-MCA compared to the reference period). MBFV diagrams for both MCA were similar with minor differences – the mean value of the difference was minimal (0.6 ± 1.85% higher in the L-MCA compared to the R-MCA). MBFV returned to baseline values in both MCA in the time period 70–75 s after task start.

Verbal fluency test: A statistically significant MBFV increase was recorded in both MCA with maximal MBFV reached in the time period 5–10 s from the task start (a 8.91 ± 3.00% increase in the R-MCA, and 8.45 ± 3.53% increase in the L-MCA compared to the reference period). Afterwards, MBFV decreased in both MCA, reaching minimal levels in the time period 20–25 s after task start. The maximal separation in MBFV between MCA was reached in the period 20–25 s when MBFV was significantly higher in the L-MCA than in the R-MCA (in 13 subjects, 2.68 ± 2.69, P = 0.013, Fig. 1a).

Figure 1.

Diagrams representing relative mean blood flow velocity (MBFV) changes in time in (image) right middle cerebral arteries (MCA) and (image) left MCA during performance of (a) phonemic verbal fluency test, (b) Stroop test with neutral stimulus, (c) Stroop test with congruent stimulus, and (d) Stroop test with incongruent stimulus. Activation period of each cognitive task is shaded in grey. Statistically significant MBFV increases for each cognitive task are marked with ‘*’. Dotted lines show the left minus right significant differences of MBFV (indicating left-sided dominance during performance of pVFT and Stroop test with incongruent stimulus).

Stroop test with neutral stimulus: a gradual MBFV increase was recorded in both MCA reaching maximal values in the time period 15–20 s after task start (a 4.26 ± 5.71% increase in the R-MCA, and a 4.47 ± 7.10% increase in the L-MCA in comparison with the reference period, both statistically significant). A maximal separation of MBFV between L-MCA and R-MCA was reached at the time period 15–20 s, MBFV being statistically insignificantly higher in the L-MCA compared to R-MCA (3.59 ± 7.75, P = 0.116, Fig. 1b).

Stroop test with congruent stimulus: a rapid statistically significant MBFV increase was recorded in the R-MCA in the time period 0–5 s (a 5.52 ± 5.56% increase compared to the reference period) and in the L-MCA in the time period 10–15 s from the task start (a 7.18 ± 7.01% increase compared to the reference period) and was followed by a gradual MBFV drop in both arteries. The maximal separation in MBFV between R-MCA and L-MCA was reached in the time period 10–15 s from the task start, MBFV being insignificantly higher in the L-MCA than in the R-MCA (in 11 subjects; in three subjects, MBFV was higher in the R-MCA; 2.53 ± 5.17%, P = 0.152, Fig. 1c).

Stroop test with incongruent stimulus: a statistically significant rapid MBFV increase was recorded in both MCA, reaching maximal values in the time period 0–5 s (a 5.44 ± 4.86% increase in R-MCA and a 5.37 ± 3.23% increase in L-MCA compared to the reference period). The increase was followed by a gradual MBFV decrease in both MCA. Maximal separation between MBFV in MCA was reached in the time period 15–20 s, MBFV being significantly higher in L-MCA than in R-MCA (in 12 subjects; in two subjects, MBFV was higher in R-MCA at the given period of time; 3.08 ± 5.32, P = 0.046, Fig. 1d).

TMTA: a rapid, statistically significant MBFV increase followed by a decrease in MBFV was recorded within the first 10 s from the task start. A bilateral, gradual MBFV increase has been recorded afterwards reaching maximal values during the time period 25–30 s after task start (a 10.07 ± 10.90% increase in the R-MCA and 8.75 ± 8.60 in L-MCA compared with the reference period). Maximal separation between MBFV in MCA was reached in the time period 5–10 s after task start, MBFV being insignificantly higher in R-MCA than in L-MCA (in 11 subjects; in three subjects, MBFV was higher in L-MCA at the given period of time; −2.48 ± 8.57, P = 0.311, Fig. 2a).

Figure 2.

Diagrams representing relative mean blood flow velocity (MBFV) changes in time in (image) right middle cerebral arteries (MCA) and (image) left MCA during performance of (a) Trail Making Test A and (b) Trail Making Test B. Activation period of each cognitive task is shaded in grey. Statistically significant MBFV increases for each cognitive task are marked with ‘*’.

TMTB: a rapid, bilateral MBFV increase was recorded in both MCA within the first 5 s followed by a gradual MBFV increase reaching maximal values in the time period 40–45 s in both MCA (a statistically significant increase of 22.59 ± 11.96% in R-MCA and of 23.44 ± 17.53 in L-MCA compared to the reference period). Maximal separation between MBFV in MCA was reached in the time period 10–15 s after task start, MBFV being insignificantly higher in R-MCA than in L-MCA (in nine subjects; in five subjects, MBFV was higher in L-MCA at the given period of time; −1.13 ± 5.28, P = 0.463, Fig. 2b).

Statistically significant L-MCA dominance was found during the performance of the Stroop test with incongruent stimulus (LI = 3.08 ± 5.32, P = 0.018), while during the performance of the pVFT no significant laterality was found (LI = 2.27 ± 0.98, P = 0.091). No significant laterality indices were found during performance of other cognitive tests. When comparing MBFV increases among cognitive tasks, the maximal MBFV increase was found during the activation phase of TMTB (23.28% ± 13.08%) and pFVT (8.92% ± 3.24%). Among two tests, the MBFV increase was significantly higher during the activation phase of TMTB (P = 0.008).

During the activation period of each cognitive task, no statistically significant correlation was found between HR and MBFV in MCA.

Discussion

The aim of our study was to obtain temporal patterns of MBFV changes in MCA during the performance of cognitive tasks (pVFT, Stroop tests and TMT). As pVFT[9-12] and BHT[24] were already investigated in TCD studies, we used these tests to validate our diagnostic system and other cognitive tests that have not been investigated in functional TCD (fTCD) studies so far (Stroop tests and TMT). A significant increase of MBFV in both MCA was found during performance of all cognitive tasks.

A rapid, nonselective MBFV increase was recorded within the first 5–10 s during performance of all cognitive tasks as well as in BHT. A similar increase was found in the study published by Szirmai et al.[11] during performance of VFT and mental arithmetic test. This MBFV increase could not be explained by metabolically induced vasodilatation of brain vessels, but rather by perfusion pressure elevation or fast neurogenic regulation of vascular resistance in whole brain vascular pool.[25] This direct neurogenic regulation possibly promotes rapid microvascular adaptation leading to significant bilateral MBFV increase in both MCA during excitement, stress and initiation of cognitive activity. Due to their lower temporal resolution, PET and fMRI would probably fail to show this rapid and transient increase of MBFV.

During BHT performance, two phases of MBFV elevation were observed. The first, rapid MBFV increase was recorded within the first 5 s of the test (stress-induced neurogenic regulation) and a second, slow increase was recorded starting 10 s afterwards and reaching maximum 10–15 s after completing the task with final decrease to normal values. The characteristic MBFV pattern of the latter phase was in accordance with previously published data.[25]

Two phases of MBFV elevation were also found during pVFT and TMTA performance. After the initial, rapid MBFV increase, a later slower MBFV increase was observed. In the Stroop test with incongruent stimulus, prolonged significant MBFV decrease, found in our study, may be due to rapid speaking and breathing during the performance of the task in highly motivated individuals trying to finish the task as quickly as possible. This may be seen especially during more demanding tasks.[11] A similar MBFV decrease was observed during voluntary hyperventilation starting approximately 10 s after starting the task, reaching minimal levels at the end of the task and finally returning to the resting levels 30 s after completing the task.[24] In contrast to our study where this decrease was approximately 15%, the magnitude of decrease was much higher during voluntary hyperventilation (40–50%), which might be due to difference in rate and depth of breathing between these two tasks.[24] The profound decrease of MBFV in the Stroop test with incongruent stimulus did not hide the laterality of flow (this is comparable to the diagram of pVFT performance). The time course of TMTB task was monophasic with rapid increase of MBFV in the first 5 s and subsequent gradual increase of MBFV during the whole task performance reaching a maximum at the end of the task. We also found a similar monophasic diagram in our anterior cerebral artery study during performance of TMTB (although significantly lower than in MCA)[26] and voluntary leg movements, but also in this study during the performance of the Stroop test with congruent stimulus.

During the pVFT's slower, later phase, an insignificant left laterality of MBFV was found showing higher MBFV in L-MCA than in R-MCA during the task performance (P = 0.091). In contrast, previous fMRI and PET[13, 14, 27, 28] as well as TCD studies[10, 11, 28, 29] have shown significant left laterality. This discrepancy might be explained by the difference in statistical analysis of laterality indices among studies. Statistically significant left lateralization was found during the performance of the Stroop test with incongruent (more demanding) stimuli, but not in the Stroop test with neutral or congruent (easy) stimulus, which confirms findings of previous studies describing less lateralization during tasks requiring well-learned automatic processing, such as reading,[12] and which was in accordance with previously functional imaging studies.[18] Our finding of no significant left–right difference during performance of TMT is in accordance with previous studies showing no hemispheric dominance and activation of both hemispheres during performance of TMT, especially TMTB.[31, 32]

In few individual cases, EHI index score and LI of some cognitive tasks showed a discrepancy that might be partly explained by bilateral representation of verbal abilities, ‘learned’ right-handedness and inclusion of right-handers with a wide range of EHI indexes. Using TCD to confirm hemispheric dominance in these cases could be unreliable.

The rapid HR increase at the beginning of most tasks found in this research confirms previous studies,[33] and is probably stress-related. Moreover, more complex cognitive tasks lead to a more significant HR increase (in our study the highest HR increase was found at the beginning of the TMTB and the Stroop test with incongruent stimulus). Additionally, our finding of no correlation between MBFV and HR is in accordance with previous studies.[33]

The methodological advantages of our study were: homogeneity of the study group regarding handedness and age; whispering during speaking-demanding tasks (pVFT and the Stroop tasks) to exclude possible artifacts due to acoustic interference with the TCD signal and to avoid masking of potential lateralization[10] due to wider bilateral cerebral presentation of loud speech.[34] On the other hand, the selection of a homogeneous group of healthy, young, right-handed volunteers limits the application of the results in the older individuals or left-handers. An additional limitation of our and comparable studies is the small sample size, which is probably caused by the complex postmonitoring analysis. To overcome these limitations, we have already broadened our research to different age groups and sexes, increasing the sample size and trying to establish the parameters that will allow us to differentiate patients with impaired cerebrovascular hemodynamics, neurodegenerative or psychiatric diseases. Hopefully, this will allow us to extend our knowledge of the association of aging, cognitive decline and dysfunction of cerebral hemodynamics and to use rapid MBFV changes as a cerebral hemodynamic marker of these entities.

In conclusion, TCD monitoring has a good potential in obtaining crude functional anatomy and specific MBFV temporal patterns of specific cognitive tasks. Some cognitive tasks (pVFT and TMTB) have shown great potential for monitoring of MBFV in MCA. TCD monitoring might be used in future neuropsychological studies (in both healthy individuals and patients with dysfunction of cerebral hemodynamics, neurodegenerative and psychiatric diseases). Additionally, to the best of our knowledge, so far, no tests used in this research (TMT and the Stroop tests) have been investigated in fTCD studies.

Acknowledgments

The study was done at the Department of Neurology, University Hospital Centre Zagreb. Authors’ contributions: Study concept and design: Boban and Malojčić. Acquisition of data: Boban, Črnac, Junaković, Malojčić. Analysis and interpretation of data: Boban, Črnac, Malojčić. Drafting of the manuscript: Boban and Črnac. Critical revision of the manuscript for important intellectual content: Boban, Malojčić, Črnac. Administrative, technical, and material support: Malojčić and Junaković. There is no conflict of interest.

Ancillary