- Email: [email protected]
Time-locked association between rapid cerebral blood flow modulation and attentional performance
Clinical Neurophysiology 119 (2008) 1292–1299 www.elsevier.com/locate/clinph
Time-locked association between rapid cerebral blood flow modulation and attentional performance S. Duscheka,*, D. Schuepbachb, R. Schandrya a
University of Munich, Department of Psychology, Leopoldstrasse 13, 80802 Mu¨nchen, Germany b Psychiatric University Hospital Zu¨rich, Lenggstrasse 31, 8032 Zu¨rich, Switzerland Accepted 16 January 2008 Available online 3 April 2008
Abstract Objective: The study investigated relationships between rapid cerebral hemodynamic modulation and attentional performance. Based on former results on complex cognitive functioning, a specific association between the first seconds of the hemodynamic response and performance was hypothesized. Methods: Using transcranial Doppler sonography, blood flow velocities in the middle cerebral arteries of both hemispheres were recorded in 48 healthy subjects. The applied task comprised motor reactions on a visual stimulus which was preceded by an acoustic warning signal (interstimulus interval 5 s). Task-induced hemodynamic changes were assessed second-by-second, and related to reaction time using analysis of variance and linear regression. Results: A right dominant blood flow response was observed. Flow velocity increase in the middle fraction of the interstimulus interval, i.e. seconds 2 and 3 after the cuing signal, significantly correlated with reaction time. This was not the case for the very early and late components of the response. Conclusions: The results suggest a time-locked association between cerebral blood flow increase and attentional performance. This is in accordance with neurophysiological studies that revealed the closest relationship between brain perfusion and cortical activity during a similar time window. Significance: The study supports the assumption of a specific, relatively early time interval in which relationships between cerebral blood flow and behavior become apparent. Ó 2008 International Federation of Clinical Neurophysiology. Published by Elsevier Ireland Ltd. All rights reserved. Keywords: Cerebral blood flow; Cerebral hemodynamic modulation; Neurovascular coupling; Attention; Doppler sonography
1. Introduction The functional coupling between neural activity and cerebral blood flow, which was first described more than a century ago (Mosso, 1881; Roy and Sherrington, 1890), is of vast importance for the investigation of brain–behavior relationships (Iadecola, 2004). Extensively used neuroimaging methods such as functional magnetic resonance imaging (fMRI), positron emission tomography (PET) or * Corresponding author. Address: Ludwig-Maximilians-Universita¨t Mu¨nchen, Department Psychologie, Leopoldstr. 13, 80802 Munich, Germany. Tel.: +49 89 2180 5297; fax: +49 89 2180 5233. E-mail address: [email protected] (S. Duschek).
single photon emission computed tomography (SPECT) make use of the cerebral hemodynamic response to map cerebral function (Logothetis et al., 2001; Iadecola, 2004). While functional brain mapping studies commonly focus on local distribution patterns of cerebral blood flow, less importance is attached to the temporal dimension of the hemodynamic response (Duschek and Schandry, 2003). This is at least partly due to the fact that the neuroimaging techniques, which provide high spatial resolution, are limited in detecting short-time changes in perfusion, and thus in their ability to effectively analyze temporal dynamics of cerebral blood flow (Stroobant and Vingerhoets, 2000). Nevertheless, high resolution dynamic analysis of blood flow is certainly valuable, for instance, when it
1388-2457/$34.00 Ó 2008 International Federation of Clinical Neurophysiology. Published by Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.clinph.2008.01.102
S. Duschek et al. / Clinical Neurophysiology 119 (2008) 1292–1299
comes to investigate neural activation processes related to fast cognitive processes like those of attention and perception (e.g. Ba¨cker et al., 1994, 1999; Knecht et al., 1996; for an overview see Duschek and Schandry, 2003). The analysis of time dynamics of cerebral blood flow has also proved helpful to gain insight into the relationship between brain perfusion and cognitive performance. A number of studies on this relationship have applied functional transcranial Doppler sonography (fTCD). This ultrasonic technique enables continuous measurement of blood flow velocities (BFV) in large cerebral arteries with excellent temporal resolution (Deppe et al., 2004). Vingerhoets and Stroobant (2002) reported that the level of the increase in BFV in the middle and anterior cerebral arteries (MCA and ACA) evoked by arithmetic processing depended on task difficulty. Duschek and Schandry (2004, 2006) monitored BFV in the MCA during the execution of attention and arithmetic tasks. Task-induced increases in BFV were significantly associated with task performance. However, the correlations were relatively low with blood flow modulations explaining not more than 10% of the variance in performance. Further fTCD studies failed to predict cognitive performance on the basis of BFV modulations in the MCA and ACA induced by tasks aiming at higher cognitive functioning (Schuepbach et al., 2002; Frauenfelder et al., 2004; Duschek et al., 2008). A very recent fTCD study on this relationship using an innovative methodological approach was presented by Schuepbach et al. (2007). They investigated BFV modulations in the MCA and ACA evoked by the ‘‘Stockings of Cambridge” (Frauenfelder et al., 2004), a computerized task on mental planning abilities. Instead of the entire hemodynamic response, their analysis focused on its initial component, i.e. the rapid increase in BFV which is usually observed during the first seconds of the execution of a cognitive task. Second-by-second analysis of the response revealed a time-locked association between early BFV modulation and task performance: The flow velocity increase in the left MCA during the 2nd second after task onset explained more than 60% of the variance in performance. No significant relationships were found in later time intervals. The authors concluded that associations between the hemodynamic response and cognitive performance become apparent in a specific and relatively early time window. They hypothesized that the connection between behavior and cerebral blood flow modulation is highly dynamic and tightly limited in time. Another line of research investigated relationships between cerebral blood flow and electrophysiological measures of neural activity. Szirmai et al. (2005) simultaneously recorded the spontaneous EEG and flow velocities in the MCA during various cognitive tasks. They found close associations between certain EEG frequencies and taskinduced modulations of BFV which were, however, restricted to tightly delimited phases of the response. In animals, the spatial distribution of local field potentials recorded from the somatosensory cortex exhibited the most
1293
precise overlap with cerebral blood flow modulation during the initial 2–3 s of the response (Sheth et al., 2005). These findings suggest that also the association between cortical activation and changes in brain perfusion is time limited, and that high resolution analysis proves useful to get further insight into this linkage. The purpose of the present study was to replicate and to extend the findings of Schuepbach et al. (2007). Considering the applied cognitive task, their conclusions are restricted to mental planning, and thus cannot be generalized to other fields of cognitive functioning. Mental planning tasks aim at highest cognitive processing, i.e. executive functioning. Hence, it seemed suggestive to investigate whether the notion of a time-locked relationship between the early cerebral hemodynamic response and performance also applies for more basic cognitive processing such as in the field of attention. Specifically, the current study focused on the arousing component of attention using a cued reaction time task. Paradigms of this type aim at preparatory processes in terms of a short-term increase of attentiveness during the anticipation of a significant event, which enables a rapid adjustment to situational requirements. This function of ‘‘phasic alertness” is considered to be a specific component of the human attentional system (Posner and Rafal, 1987; Johnson and Proctor, 2004) that is undoubtedly of vast importance in everyday life. At the cerebral level, the arousing component of attention is represented by cortical as well as subcortical structures (Johnson and Proctor, 2004; Sturm et al., 1999). Aside from the reticular formation, brain imaging studies have shown that the anterior cingulate, as well as the dorsolateral frontal and the inferior parietal lobes, are of great importance for this cognitive function (e.g. Paus et al., 1997). A right hemispheric dominance of this network is assumed (Posner and Petersen, 1990; Sturm et al., 1999). In the present study, the relationship between rapid BFV modulations in the MCA and cognitive performance was investigated based on a cued reaction time task. The hemodynamic response during task execution was recorded using fTCD, and quantified in consecutive time windows of 1 s duration. Considering the findings of Schuepbach et al. (2007) and Sheth et al. (2005), BFV changes in the early phase of the response, i.e. seconds 2–3 after task onset, were expected to best predict reaction time. 2. Methods 2.1. Participants Forty-eight subjects (11 men, 37 women) participated in the study. Exclusion criteria comprised severe physical diseases, psychiatric disorders, as well as the use of psychoactive drugs, analgesics or medication affecting the cardiovascular system. All participants were right-handed according to the Edinburgh Handedness Inventory (Oldfield, 1971) (the mean laterality quotient was 89.8, the
1294
S. Duschek et al. / Clinical Neurophysiology 119 (2008) 1292–1299
Table 1 Age, Body Mass Index, blood pressure and MCA resting blood flow velocities in the sample
Age in years Body Mass Index in kg/m2 Systolic blood pressure in mmHg Diastolic blood pressure in mmHg Resting flow velocity left MCA in cm/s Resting flow velocity right MCA in cm/s
M
SD
Min
Max
29.3 21.7 117.1 74.7 67.3 66.1
7.1 2.2 8.4 7.8 11.2 11.1
20 17.7 100 58 43.5 41.9
52 28.1 137 91 93.9 92.1
Means (M), standard deviations (SD), minimal (Min) and maximal (Max) values.
pressure was measured using an automatic inflation blood pressure monitor (MIT, TYP M CR15; Omron, USA). After the mounting of the ultrasonic probes, flow velocities in both MCAs were recorded during a rest period of 3 min. For this purpose, participants were asked not to speak and to relax with their eyes opened. Following this, the cued reaction time task was presented in the described form. Subjects were requested not to drink alcohol or beverages containing caffeine for three hours prior to the experimental session. 2.5. Data analysis
range being 80–100). Thirty-seven of the participants were university students, eight were employees and three were self-employed. Information regarding age, Body Mass Index (BMI, kg/m2), as well as blood pressure and resting BFV in the left and right MCAs (see Section 2.4) is presented in Table 1. 2.2. Task characteristics The task was presented on a computer screen using the ‘‘Experimental Runtime System” software program (BeriSoft Cooperation, 2000). On the screen, the white outline of a small cross (6 6 mm) was shown. After 55 s, an acoustic cue was presented (400-Hz tone of 500 ms duration). Five seconds after the cue, the image was replaced by a full white cross of the same size. This served as the imperative stimulus requiring an immediate keystroke. The task consisted of a total of 20 trials. In order to control for laterality effects, half of the participants carried out the first 10 trials with the right hand and the remaining with the left hand. The sequence was reversed in the second half of the participants. The subjects were asked to sit still, not to speak and to look at the cross during the entire duration of the task. The reaction times were recorded automatically and aggregated by calculating the median for each subject.
2.5.1. Analysis of the hemodynamic data The envelope curves revealed by Doppler sonography were analyzed offline using the software AVERAGE (Deppe et al., 1997). The BFV data recorded during task execution were integrated over each cardiac cycle and averaged, time-locked to the cuing tone. For each trial, a time segment was extracted beginning 10 s before the cuing tone and ending 1 s after the imperative stimulus. In order to minimize artifacts, trials containing values below 60% or above 150% of the mean BFV recording of the respective subject were rejected. According to this procedure, an average of 0.2 of the 20 trials (SD = 1.0) per subject had to be eliminated. The mean BFV during the 10 s prior to the cuing tone served as a baseline (BFVbas). Relative changes in BFV during task execution (dBFV) were calculated for the left and the right MCAs using the formula dBFV = [BFV(t)BFVbas] 100/BFVbas, where BFV(t) is the flow velocity over the course of time. Following this, the mean values of dBFV were computed for 6 response intervals of 1 s duration each. The response intervals included the 5 s of the interstimulus interval, as well as the 1st second after the imperative stimulus. During the 1st second following the imperative stimulus all subjects executed the motor responses (reaction time range 206–430 ms). Thus, the 6 response intervals covered the entire duration of the task.
2.3. Recording of BFV A commercially available transcranial Doppler sonography device (Multidop L2, DWL Elektronische Systeme) was employed. BFV recordings were accomplished simultaneously in the left and right MCAs. The recordings were obtained through the temporal bone windows, using two 2-MHz transducer probes. The MCAs were insonated at a depth of 50 mm in all subjects. Following vessel identification, the ultrasonic probes were fixed to the head using a tight rubber band. The spectral envelope curves of the Doppler signal were stored at a rate of 28 samples per second. 2.4. Procedure The experimental sessions were conducted in a silent, dimly lit room. Prior to the experimental procedure, blood
2.5.2. Statistical analysis The task-induced changes in MCA flow velocities and their relationship with reaction time were evaluated using a three-way repeated measures ANOVA. Prior to this analysis, the sample was split according to the reaction time median of the sample (294.8 ms). The resulting performance groups (i.e. high vs. low reaction time) entered the model as between-subjects factor. Time (i.e. the 6 response intervals) and hemisphere (i.e. left MCA vs. right MCA) were included as within-subject factors. The values of dBFV for the 6 response intervals served as a dependent variable. The performance groups did not differ with regard to age (t = 0.56, p = .58), gender (U = 252, p = .31), BMI (t = 1.47, p = .15) and blood pressure (systolic: t = 1.41, p = .17; diastolic: t = 1.31, p = .20). This also held true for resting BFV in the left (t = 1.43, p = .16) and right
(t = 0.58, p = .56) MCAs. The mean reaction time was 262.9 ms (SD = 27.9) in the high and 346.9 ms (SD = 32.3) in the low performance group (t = 9.65, p < .001). The low performance group comprised 7 men and 17 women, while the high performance group comprised 4 men and 20 women. No significant differences between men and women were found with respect to reaction time (men: M = 288.3 ms, SD = 57.6 ms; women: M = 309.9, SD = 49.8, t = 1.22, p = .23) and the dBFV values for each of the 6 response intervals (all p > .10). Post hoc t-tests were applied to determine more precisely in which of the response intervals flow velocities in the MCA exceeded baseline level. For this purpose, single sample t-tests were used to test whether the dBFV values significantly differed from zero. In order to account for multiple testing (12 comparisons), alpha was set at .004. The 12 dBFV values were furthermore compared between both performance groups (alpha = .004). Finally, the values for the 6 response intervals were compared between both hemispheres (6 comparisons, alpha = .008). To complement the categorical analysis, stepwise multiple regression analyses were employed. The dBFV values for the 6 response intervals served as predictors and reaction time as dependant variable (alpha for t = .004). Two separate models were computed referring to the left and right MCAs. 3. Results Fig. 1 displays the time course of the BFV changes in the MCAs during the execution of the cued reaction time task. A steep bilateral increase of BFV was observed with a latency of less than 1 s after the cuing tone. A maximum appeared approximately 3.5 s after this stimulus followed by a continuous decline until the end of the registration period. The BFV increase was stronger in the right than in the left MCA beginning at second 2 after the cuing tone until the end of the registration period. In Fig. 2, the dBFV values for the 6 response intervals are shown for both performance groups. Except for the 1st second, the values were consistently higher in the high performance group. This was the case for both the left and right MCAs. The ANOVA revealed significant main effects for time (F[5, 230] = 64.99, p < .001; partial Eta squared = .59) and hemisphere (F[1, 46] = 7.82, p = .008; partial Eta squared = .15) which corroborated the overall taskinduced change in MCA flow velocities, as well as the right-hemispheric lateralization of the response. The factor performance group was also significant (F[1, 46] = 8.72, p = .005; partial Eta squared = .16) confirming the overall higher dBFV values in subjects with shorter reaction times. Furthermore, a significant interaction between the factors time and performance group was found (F[5, 230] = 3.57, p = .004; partial Eta squared = .072).
Change in flow velocity in %
S. Duschek et al. / Clinical Neurophysiology 119 (2008) 1292–1299
1295
Imperative stimulus
5 4 3 2
Cuing tone
Left MCA
1
Right MCA
0 -1
0
1
2
3
4
5
6
Time in s Fig. 1. Changes in MCA blood flow velocities during the execution of the cued reaction time task (grand average).
This indicated that the size of the group differences varied across the 6 time intervals. Time significantly interacted with hemisphere (F[5, 230] = 14.71, p < .001; partial Eta squared = .24) corroborating the stronger lateralization of the later as compared to the earlier phase of the hemodynamic response. The results of post hoc testing are given in Table 2. The dBFV values for the left and right MCAs differed significantly from zero (baseline level) in the case of each response interval except for the 1st second. In the seconds 2 and 3, significantly higher dBFV values for both vessels were found in the high performance group. In the right MCA, a significant group difference also emerged in second 4. The hemispheric difference was only significant in seconds 5 and 6. Table 3 displays the results of multiple regression analysis. Only the dBFV values for the third second significantly predicted reaction time. This held true for the models referring to both the left and the right MCAs. The dBFV values concerning the other response intervals were excluded in the first step, and no further models reached significance. 4. Discussion Among the brain areas relevant for the control of attentional arousal, the dorsolateral frontal and the inferior parietal lobes are parts of the perfusion territory of the MCA (Pardo et al., 1991; Paus et al., 1997; Haines, 2007). It seems plausible to ascribe the hemodynamic response observed during the execution of the reaction time task to neural activation processes in these structures. The more pronounced BFV increase in the right MCA emphasizes the dominance of the right hemisphere for arousal regulation (Posner and Petersen, 1990; Sturm et al., 1999).
1296
S. Duschek et al. / Clinical Neurophysiology 119 (2008) 1292–1299
Left MCA High performance Low performance
10
Change in flow velocity in %
Right MCA High performance Low performance
10
*
8
8
*
* 6
6
4
4
*
*
2
2
0
0
-2
-2 1
2
3
4
5
6
1
2
3
4
5
6
Time in s Fig. 2. Values of dBFV for both performance groups. Seconds 1–5 include the interstimulus interval. In second 6, the response was executed (bars represent standard deviations,* for p < .004).
The linkage between neural processes and cerebral hemodynamics is mediated by various physiological and biochemical mechanisms. At this, the tight coupling between nerve-cell activity and brain metabolism may play the most prominent role (Paulson, 2002; Iadecola, 2004). Controlled by a variety of neurochemical factors like K+, H+ and adenosine, neural activation leads to the dilation of cerebral arterioles and capillaries followed by an increase in local cerebral blood flow (Iadecola, 2004). Besides flow metabolism coupling, fast acting neural mechanisms contribute to cerebrovascular regulation. An intracranial vasodilative system is assumed which directly triggers the dilation of cortical microvessels as a response to brain stem activation (Sa´ndor, 1999; Sato et al., 2001). In animals, for instance, the experimental stimulation of
the Nucleus basalis Meynert of the brain stem was shown to produce cortical blood flow increase (Biesold et al., 1989). This mechanism is mediated by afferent fibres connecting lower brain areas with cortical microvessels (c.f. Sa´ndor, 1999; Hamel et al., 2002). In particular, cholinergic and serotonergic neurons are involved originating in the ascending reticular activating system (e.g. Sato et al., 2001; Szirmai et al., 2005). The ascending reticular activating system is of vast importance for the control of attentional arousal (Posner and Petersen, 1990), thus it is suggestive to assume an important role of the described neural system in the context of the present task. This rapid mechanism may be particularly involved in mediating the steep initial increase of BFV, which was observed with a latency of less than one
Table 2 Post hoc t-tests for differences of the dBFV values from zero (baseline), as well as for differences in the dBFV values between both performance groups and between both hemispheres Response interval
Baseline vs. task execution
High vs. low performance
*
*
t ( for p < .004)
t ( for p < .004)
Left vs. right MCA t (* for p < .008)
Second 1
Left MCA Right MCA
1.51 1.23
Left MCA Right MCA
0.58 0.66
0.57
Second 2
Left MCA Right MCA
3.03* 3.29*
Left MCA Right MCA
2.81* 2.81*
0.37
Second 3
Left MCA Right MCA
8.33* 9.28*
Left MCA Right MCA
3.15* 3.31*
1.72
Second 4
Left MCA Right MCA
8.49* 9.37*
Left MCA Right MCA
2.77 3.18*
2.34
Second 5
Left MCA Right MCA
7.48* 8.93*
Left MCA Right MCA
2.25 2.60
3.28*
Second 6
Left MCA Right MCA
5.99* 8.00*
Left MCA Right MCA
1.73 2.33
5.00*
S. Duschek et al. / Clinical Neurophysiology 119 (2008) 1292–1299 Table 3 Regression analyses for the prediction of reaction time from the dBFV values for the left and right MCAs, Beta and t-values (left MCA: R2 = .19, right MCA: R2 = .22) Response interval
Left MCA Beta
Second Second Second Second Second Second
1 2 3 4 5 6
.15 .12 .43 .05 .08 .13
*
t ( for p < .004) 1.03 0.48 3.23* 0.14 0.35 0.77
Right MCA Beta .16 .26 .47 .24 .02 .04
t (* for p < .004) 1.11 1.03 3.64* 0.71 0.07 0.24
second after the cuing signal. The contribution of both metabolic and non-metabolic factors to the hemodynamic response is also supported by the time-course of its lateralization across the task period. Given the higher metabolic activity in the right hemisphere related to preparatory attention (Sturm et al., 1999), flow metabolism coupling may be relevant mainly in the later phase of the response in which marked lateralization occurred. In contrast, non-metabolic factors may be dominant in triggering the non-lateralized initial response. Also an involvement of systemic blood pressure in the development of the cerebral hemodynamic response was reported (for overview see Duschek and Schandry, 2007). Autoregulatory processes are commonly assumed to keep cerebral blood flow constant by buffering transient blood pressure fluctuations (Chillon and Baumbach, 1997; Paulson, 2002). Nonetheless, the extent of cerebral hemodynamic modulation during mental activity has been shown to be moderated both by resting blood pressure and by cognitively induced blood pressure changes (Duschek and Schandry, 2004, 2006). Interestingly, the effects of blood pressure and heart rate on cerebral hemodynamic modulation during mental activity were also found to be most pronounced during the first seconds of the response (Duschek et al., 2008). In the present study, the task-induced increase in MCA flow velocities was substantially more pronounced in individuals who exhibited shorter reaction times. The interaction between the factors time and performance suggested differently pronounced effects of specific components of the hemodynamic response on reaction time. Post hoc testing based on an adjusted significance criterion indicated that the performance groups significantly differed with respect to the increase of BFV in the left MCA during seconds 2 and 3 after task onset. For the right MCA, a significant group difference was also found in second 4. Modulation of BFV in seconds 1, 5 and 6 did not differ significantly between groups. It is evident that the middle fraction of the hemodynamic response proved to predict reaction speed more precisely than did the later response component during which the motor reaction was executed. This is also supported by regression analysis which revealed that BFV modulation in both MCAs significantly predicted task performance only during the third second.
1297
The amounts of shared variance between BFV modulation and reaction time were 19% and 22% for the left and right vessels, respectively. This suggests a markedly closer relationship as compared to earlier fTCD studies using conventional analysis of hemodynamic modulation. Employing the same reaction time task, but averaging the BFV response in the MCAs across the entire interstimulus interval, yielded not more than 4% shared variance with reaction time (Duschek and Schandry, 2004). In a study based on a serial subtraction task, 9% common variance between overall task-induced increase in MCA flow velocities and arithmetic performance was found (Duschek and Schandry, 2006). The present finding is in line with the report of Schuepbach et al. (2007) on the relationship between BFV modulation and performance in mental planning. Schuepbach et al. (2007), however, found the closest association between MCA flow velocity modulation and task performance in an even earlier time window, i.e. the 2nd second after task onset. The present categorical analysis furthermore suggested a somewhat longer duration of the relevant time interval, namely 2–3 s. As an explanation of the divergent results, different time courses of the hemodynamic responses related to the varying cognitive requirements of the applied tasks may be considered. Schuepbach et al. (2007) observed a bilateral BFV maximum in second 2 of the 5-second interval during which mental planning took place. In the reaction time task, BFV peaked somewhat later, i.e. in second 4. One may speculate that this slower and longer lasting ascent of BFV was due to a continuous increase in attentional arousal during the preparation for the imperative stimulus. The higher magnitude of the later response component, in turn, could explain its closer relationship with performance. Nonetheless, the results of both studies are consistent in the sense that both clearly support the assumption of a specific and relatively early time window in which close relationships between BFV modulation and cognitive performance become apparent. The present study underlines that this notion is not restricted to executive functioning, but also applies to more basic cognitive processes such as preparatory attention. Also studies investigating cerebral blood flow using methods other than fTCD corroborated the specific relevance of the early hemodynamic response. For instance, volume-based hemodynamic measurements by means of optical imaging and laser Doppler flowmetry revealed a more precise spatial overlap with cortical field potentials during the first seconds of sensory stimulation than in later time intervals (Sheth et al., 2005). Functional brain imaging based on echoplanar MRI and gradient echo MRI suggested that ocular dominance columns may be most accurately resolved using the first few seconds of the BOLD response (Menon and Goodyear, 1999; Duong et al., 2000). Thus, also hemodynamic measurements based on blood volume and oxygenation underline the importance of the temporal dimension of cerebral blood flow modulation
1298
S. Duschek et al. / Clinical Neurophysiology 119 (2008) 1292–1299
for the investigation of brain behavior relationships (Sheth et al., 2005). One limitation of the study results from the restriction of the BFV recordings to the MCAs. Some of the brain areas relevant for the control of attentional arousal are not part of their perfusion territory. This holds particularly true for the anterior cingulate which is supplied by branches of the ACAs (Sturm et al., 1999; Haines, 2007). Activation processes in this structure are inevitably not reflected by the present measurement. In conclusion, the current study adds evidence to the notion of a time-locked relationship between rapid cerebral blood flow modulation and cognitive performance. In accordance with neurophysiological findings, the available fTCD studies suggest the existence of a time window of 2– 3 s after activation in which the cerebral hemodynamic response exhibits the closest linkage with neural activation, as well as basic and complex behavior. Future neurophysiological research in this area could certainly benefit from a stronger focus on the dynamic aspects of cerebral blood flow in addition to local hemodynamic distribution patterns. Acknowledgements We are grateful to Alessandro Angrilli and Ann-Kristin Adam for their help with the analysis of the hemodynamic data. References Ba¨cker M, Knecht S, Deppe M, Lohmann H, Ringelstein EB, Henningsen H. Cortical tuning: a function of anticipated stimulus intensity. Neuroreport 1999;10:293–6. Ba¨cker M, Deppe M, Zunker P, Henningsen H, Knecht S. Tuning to somatosensory stimuli during focal attention. Cerebrovasc Dis 1994;4(Suppl. 3):3. BeriSoft Cooperation. Experimental Run Time System, Version 3.32. Frankfurt a.M.: BeriSoft Cooperation, 2000. Biesold D, Inanami O, Sato A, Sato Y. Stimulation of the nucleus basalis of Meynert increases cerebral cortical blood flow in rats. Neurosci Lett 1989;98:39–44. Chillon JM, Baumbach GL. Autoregulation of cerebral blood flow. In: Welch KMA, Caplan LR, Reis DJ, Siesjo¨ BK, Weir B, editors. Primer on Cerebrovascular Diseases. San Diego: Academic Press; 1997. p. 51–4. Deppe M, Knecht S, Henningsen H, Ringelstein EB. AVERAGE: a Windows program for automated analysis of event related cerebral blood flow. J Neurosci Meth 1997;75:147–54. Deppe M, Ringelstein EB, Knecht S. The investigation of functional brain lateralization by transcranial Doppler sonography. Neuroimage 2004;21:1124–46. Duong TQ, Kim DS, Ugurbil K, Kim SG. Spatiotemporal dynamics of the BOLD fMRI signals: toward mapping submillimeter cortical columns using the early negative response. Magn Reson Med 2000;44:231–42. Duschek S, Schandry R. Functional transcranial Doppler sonography as a tool in psychophysiological research. Psychophysiology 2003;40:436–54. Duschek S, Schandry R. Cognitive performance and cerebral blood flow in essential hypotension. Psychophysiology 2004;41:905–13.
Duschek S, Schandry R. Deficient adjustment of cerebral blood flow to cognitive activity due to chronically low blood pressure. Biol Psychol 2006;72:311–7. Duschek S, Schandry R. Reduced brain perfusion and cognitive performance due to essential hypotension. Clin Auton Res 2007;17:69–76. Duschek S, Werner N, Kapan N, Reyes del Paso GA. Patterns of cerebral blood flow and systemic hemodynamics during arithmetic processing. J Psychophysiol 2008, in press. Frauenfelder BA, Schuepbach D, Baumgartner RW, Hell D. Specific alterations of cerebral hemodynamics during a planning task: a transcranial Doppler sonography study. Neuroimage 2004;22:1223–30. Haines DE. Neuroanatomy. An Atlas of Structures, Sections, and Systems. Philadelphia: Lippincott Williams & Wilkins; 2007. Hamel E, Vaucher E, Tong XK, St-Georges M. Neuronal messengers as mediators of microvascular toe in the cerebral cortex. ICS 2002;1235:262–76. Iadecola C. Neurovascular regulation in the normal brain and in Alzheimer’s disease. Nat Rev Neurosci 2004;5:347–60. Johnson A, Proctor RW. Attention, Theory and Practice. Thousand Oaks: Sage Publications; 2004. Knecht S, Henningsen H, Deppe M, Huber T, Ebner A, Ringelstein EB. Successive activation of both cerebral hemispheres during cued word generation. Neuroreport 1996;7:820–4. Logothetis NK, Pauls J, Trinath T, Oeltermann A. Neurophysiological investigation of the basis of the fMRI signal. Nature 2001;412:150–7. Menon RS, Goodyear BG. Submillimeter functional localization in human striate cortex using BOLD contrast at 4 Tesla: implications for the vascular point-spread function. Magn Resonan Med 1999;41:230–5. ¨ ber den Kreislauf des Blutes im menschlichen Gehirn. Leipzig: Mosso A. U Von Veit; 1881. Oldfield RC. The assessment and analysis of handedness: the Edinburgh inventory. Neuropsychologia 1971;9:97–113. Pardo JV, Fox PT, Raichle ME. Localisation of a human system for sustained attention by positron emission tomography. Nature 1991;349:61–4. Paulson OB. Blood–brain barrier, brain metabolism and cerebral blood flow. Eur Neuropsychopharm 2002;12:495–501. Paus T, Zatorre RJ, Hofle N, Caramanos Z, Gorman J, Petrides M, et al. Time-related changes in neural systems underlying attention and arousal during the performance of an auditory vigilance task. J Cogn Neurosci 1997;9:392–408. Posner MI, Rafal RD. Cognitive theories of attention and the rehabilitation of attentional deficits. In: Meier MJ, Benton AL, Diller L, editors. Neuropsychological Rehabilitation. Edinburgh: Churchill Livingston; 1987. p. 182–201. Posner MI, Petersen SE. The attention system of the human brain. Annu Rev Neurosci 1990;13:25–42. Roy C, Sherrington C. On the regulation of the blood supply of the brain. J Physiol 1890;11:85–108. Sa´ndor P. Nervous control of the cerebrovascular system: doubts and facts. Neurochem Int 1999;35:237–59. Sato A, Sato Y, Uchida S. Regulation of regional cerebral blood flow by cholinergic fibres originating in the basal forebrain. Int J Dev Neurosci 2001;19:327–37. Schuepbach D, Merlo MCG, Goenner F, Staikov I, Mattle HP, Dierks T, et al. Cerebral hemodynamic response induced by the tower of Hanoi puzzle and the Wisconsin card sorting test. Neuropsychologia 2002;40:39–53. Schuepbach D, Boeker H, Duschek S, Hell D. Rapid cerebral hemodynamic modulation during mental planning and movement execution: evidence of time-locked relationship with complex behavior. Clin Neurophysiol 2007;118:2254–62. Sheth SA, Nemoto M, Guiou MW, Walker MA, Toga AW. Spatiotemporal evolution of functional hemodynamic changes and their relationship to neuronal activity. J Cerebr Blood F Metab 2005;25:830–41.
S. Duschek et al. / Clinical Neurophysiology 119 (2008) 1292–1299 Stroobant N, Vingerhoets G. Transcranial Doppler ultrasonography monitoring of cerebral hemodynamics during performance of cognitive tasks: a review. Neuropsychol Rev 2000;10:213–31. Sturm W, de Simone A, Krause BJ, Specht K, Hesselmann V, Radermacher I, et al. Functional anatomy of intrinsic alertness: evidence for a fronto-parietal-thalamic-brainstem network in the right hemisphere. Neuropsychologia 1999;37:797–805.
1299
Szirmai I, Amrein I, Pa´lvo¨gyi L, Debreczeni R, Kamondi A. Correlation between blood flow velocity in the middle cerebral artery and EEG during cognitive effort. Cogn Brain Res 2005;24:33–40. Vingerhoets G, Stroobant N. Reliability and validity of day-to-day blood flow velocity reactivity in a single subject: an fTCD study. Ultrasound Med Biol 2002;28:197–202.
Time-locked association between rapid cerebral blood flow modulation and attentional performance S. Duscheka,*, D. Schuepbachb, R. Schandrya a
University of Munich, Department of Psychology, Leopoldstrasse 13, 80802 Mu¨nchen, Germany b Psychiatric University Hospital Zu¨rich, Lenggstrasse 31, 8032 Zu¨rich, Switzerland Accepted 16 January 2008 Available online 3 April 2008
Abstract Objective: The study investigated relationships between rapid cerebral hemodynamic modulation and attentional performance. Based on former results on complex cognitive functioning, a specific association between the first seconds of the hemodynamic response and performance was hypothesized. Methods: Using transcranial Doppler sonography, blood flow velocities in the middle cerebral arteries of both hemispheres were recorded in 48 healthy subjects. The applied task comprised motor reactions on a visual stimulus which was preceded by an acoustic warning signal (interstimulus interval 5 s). Task-induced hemodynamic changes were assessed second-by-second, and related to reaction time using analysis of variance and linear regression. Results: A right dominant blood flow response was observed. Flow velocity increase in the middle fraction of the interstimulus interval, i.e. seconds 2 and 3 after the cuing signal, significantly correlated with reaction time. This was not the case for the very early and late components of the response. Conclusions: The results suggest a time-locked association between cerebral blood flow increase and attentional performance. This is in accordance with neurophysiological studies that revealed the closest relationship between brain perfusion and cortical activity during a similar time window. Significance: The study supports the assumption of a specific, relatively early time interval in which relationships between cerebral blood flow and behavior become apparent. Ó 2008 International Federation of Clinical Neurophysiology. Published by Elsevier Ireland Ltd. All rights reserved. Keywords: Cerebral blood flow; Cerebral hemodynamic modulation; Neurovascular coupling; Attention; Doppler sonography
1. Introduction The functional coupling between neural activity and cerebral blood flow, which was first described more than a century ago (Mosso, 1881; Roy and Sherrington, 1890), is of vast importance for the investigation of brain–behavior relationships (Iadecola, 2004). Extensively used neuroimaging methods such as functional magnetic resonance imaging (fMRI), positron emission tomography (PET) or * Corresponding author. Address: Ludwig-Maximilians-Universita¨t Mu¨nchen, Department Psychologie, Leopoldstr. 13, 80802 Munich, Germany. Tel.: +49 89 2180 5297; fax: +49 89 2180 5233. E-mail address: [email protected] (S. Duschek).
single photon emission computed tomography (SPECT) make use of the cerebral hemodynamic response to map cerebral function (Logothetis et al., 2001; Iadecola, 2004). While functional brain mapping studies commonly focus on local distribution patterns of cerebral blood flow, less importance is attached to the temporal dimension of the hemodynamic response (Duschek and Schandry, 2003). This is at least partly due to the fact that the neuroimaging techniques, which provide high spatial resolution, are limited in detecting short-time changes in perfusion, and thus in their ability to effectively analyze temporal dynamics of cerebral blood flow (Stroobant and Vingerhoets, 2000). Nevertheless, high resolution dynamic analysis of blood flow is certainly valuable, for instance, when it
1388-2457/$34.00 Ó 2008 International Federation of Clinical Neurophysiology. Published by Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.clinph.2008.01.102
S. Duschek et al. / Clinical Neurophysiology 119 (2008) 1292–1299
comes to investigate neural activation processes related to fast cognitive processes like those of attention and perception (e.g. Ba¨cker et al., 1994, 1999; Knecht et al., 1996; for an overview see Duschek and Schandry, 2003). The analysis of time dynamics of cerebral blood flow has also proved helpful to gain insight into the relationship between brain perfusion and cognitive performance. A number of studies on this relationship have applied functional transcranial Doppler sonography (fTCD). This ultrasonic technique enables continuous measurement of blood flow velocities (BFV) in large cerebral arteries with excellent temporal resolution (Deppe et al., 2004). Vingerhoets and Stroobant (2002) reported that the level of the increase in BFV in the middle and anterior cerebral arteries (MCA and ACA) evoked by arithmetic processing depended on task difficulty. Duschek and Schandry (2004, 2006) monitored BFV in the MCA during the execution of attention and arithmetic tasks. Task-induced increases in BFV were significantly associated with task performance. However, the correlations were relatively low with blood flow modulations explaining not more than 10% of the variance in performance. Further fTCD studies failed to predict cognitive performance on the basis of BFV modulations in the MCA and ACA induced by tasks aiming at higher cognitive functioning (Schuepbach et al., 2002; Frauenfelder et al., 2004; Duschek et al., 2008). A very recent fTCD study on this relationship using an innovative methodological approach was presented by Schuepbach et al. (2007). They investigated BFV modulations in the MCA and ACA evoked by the ‘‘Stockings of Cambridge” (Frauenfelder et al., 2004), a computerized task on mental planning abilities. Instead of the entire hemodynamic response, their analysis focused on its initial component, i.e. the rapid increase in BFV which is usually observed during the first seconds of the execution of a cognitive task. Second-by-second analysis of the response revealed a time-locked association between early BFV modulation and task performance: The flow velocity increase in the left MCA during the 2nd second after task onset explained more than 60% of the variance in performance. No significant relationships were found in later time intervals. The authors concluded that associations between the hemodynamic response and cognitive performance become apparent in a specific and relatively early time window. They hypothesized that the connection between behavior and cerebral blood flow modulation is highly dynamic and tightly limited in time. Another line of research investigated relationships between cerebral blood flow and electrophysiological measures of neural activity. Szirmai et al. (2005) simultaneously recorded the spontaneous EEG and flow velocities in the MCA during various cognitive tasks. They found close associations between certain EEG frequencies and taskinduced modulations of BFV which were, however, restricted to tightly delimited phases of the response. In animals, the spatial distribution of local field potentials recorded from the somatosensory cortex exhibited the most
1293
precise overlap with cerebral blood flow modulation during the initial 2–3 s of the response (Sheth et al., 2005). These findings suggest that also the association between cortical activation and changes in brain perfusion is time limited, and that high resolution analysis proves useful to get further insight into this linkage. The purpose of the present study was to replicate and to extend the findings of Schuepbach et al. (2007). Considering the applied cognitive task, their conclusions are restricted to mental planning, and thus cannot be generalized to other fields of cognitive functioning. Mental planning tasks aim at highest cognitive processing, i.e. executive functioning. Hence, it seemed suggestive to investigate whether the notion of a time-locked relationship between the early cerebral hemodynamic response and performance also applies for more basic cognitive processing such as in the field of attention. Specifically, the current study focused on the arousing component of attention using a cued reaction time task. Paradigms of this type aim at preparatory processes in terms of a short-term increase of attentiveness during the anticipation of a significant event, which enables a rapid adjustment to situational requirements. This function of ‘‘phasic alertness” is considered to be a specific component of the human attentional system (Posner and Rafal, 1987; Johnson and Proctor, 2004) that is undoubtedly of vast importance in everyday life. At the cerebral level, the arousing component of attention is represented by cortical as well as subcortical structures (Johnson and Proctor, 2004; Sturm et al., 1999). Aside from the reticular formation, brain imaging studies have shown that the anterior cingulate, as well as the dorsolateral frontal and the inferior parietal lobes, are of great importance for this cognitive function (e.g. Paus et al., 1997). A right hemispheric dominance of this network is assumed (Posner and Petersen, 1990; Sturm et al., 1999). In the present study, the relationship between rapid BFV modulations in the MCA and cognitive performance was investigated based on a cued reaction time task. The hemodynamic response during task execution was recorded using fTCD, and quantified in consecutive time windows of 1 s duration. Considering the findings of Schuepbach et al. (2007) and Sheth et al. (2005), BFV changes in the early phase of the response, i.e. seconds 2–3 after task onset, were expected to best predict reaction time. 2. Methods 2.1. Participants Forty-eight subjects (11 men, 37 women) participated in the study. Exclusion criteria comprised severe physical diseases, psychiatric disorders, as well as the use of psychoactive drugs, analgesics or medication affecting the cardiovascular system. All participants were right-handed according to the Edinburgh Handedness Inventory (Oldfield, 1971) (the mean laterality quotient was 89.8, the
1294
S. Duschek et al. / Clinical Neurophysiology 119 (2008) 1292–1299
Table 1 Age, Body Mass Index, blood pressure and MCA resting blood flow velocities in the sample
Age in years Body Mass Index in kg/m2 Systolic blood pressure in mmHg Diastolic blood pressure in mmHg Resting flow velocity left MCA in cm/s Resting flow velocity right MCA in cm/s
M
SD
Min
Max
29.3 21.7 117.1 74.7 67.3 66.1
7.1 2.2 8.4 7.8 11.2 11.1
20 17.7 100 58 43.5 41.9
52 28.1 137 91 93.9 92.1
Means (M), standard deviations (SD), minimal (Min) and maximal (Max) values.
pressure was measured using an automatic inflation blood pressure monitor (MIT, TYP M CR15; Omron, USA). After the mounting of the ultrasonic probes, flow velocities in both MCAs were recorded during a rest period of 3 min. For this purpose, participants were asked not to speak and to relax with their eyes opened. Following this, the cued reaction time task was presented in the described form. Subjects were requested not to drink alcohol or beverages containing caffeine for three hours prior to the experimental session. 2.5. Data analysis
range being 80–100). Thirty-seven of the participants were university students, eight were employees and three were self-employed. Information regarding age, Body Mass Index (BMI, kg/m2), as well as blood pressure and resting BFV in the left and right MCAs (see Section 2.4) is presented in Table 1. 2.2. Task characteristics The task was presented on a computer screen using the ‘‘Experimental Runtime System” software program (BeriSoft Cooperation, 2000). On the screen, the white outline of a small cross (6 6 mm) was shown. After 55 s, an acoustic cue was presented (400-Hz tone of 500 ms duration). Five seconds after the cue, the image was replaced by a full white cross of the same size. This served as the imperative stimulus requiring an immediate keystroke. The task consisted of a total of 20 trials. In order to control for laterality effects, half of the participants carried out the first 10 trials with the right hand and the remaining with the left hand. The sequence was reversed in the second half of the participants. The subjects were asked to sit still, not to speak and to look at the cross during the entire duration of the task. The reaction times were recorded automatically and aggregated by calculating the median for each subject.
2.5.1. Analysis of the hemodynamic data The envelope curves revealed by Doppler sonography were analyzed offline using the software AVERAGE (Deppe et al., 1997). The BFV data recorded during task execution were integrated over each cardiac cycle and averaged, time-locked to the cuing tone. For each trial, a time segment was extracted beginning 10 s before the cuing tone and ending 1 s after the imperative stimulus. In order to minimize artifacts, trials containing values below 60% or above 150% of the mean BFV recording of the respective subject were rejected. According to this procedure, an average of 0.2 of the 20 trials (SD = 1.0) per subject had to be eliminated. The mean BFV during the 10 s prior to the cuing tone served as a baseline (BFVbas). Relative changes in BFV during task execution (dBFV) were calculated for the left and the right MCAs using the formula dBFV = [BFV(t)BFVbas] 100/BFVbas, where BFV(t) is the flow velocity over the course of time. Following this, the mean values of dBFV were computed for 6 response intervals of 1 s duration each. The response intervals included the 5 s of the interstimulus interval, as well as the 1st second after the imperative stimulus. During the 1st second following the imperative stimulus all subjects executed the motor responses (reaction time range 206–430 ms). Thus, the 6 response intervals covered the entire duration of the task.
2.3. Recording of BFV A commercially available transcranial Doppler sonography device (Multidop L2, DWL Elektronische Systeme) was employed. BFV recordings were accomplished simultaneously in the left and right MCAs. The recordings were obtained through the temporal bone windows, using two 2-MHz transducer probes. The MCAs were insonated at a depth of 50 mm in all subjects. Following vessel identification, the ultrasonic probes were fixed to the head using a tight rubber band. The spectral envelope curves of the Doppler signal were stored at a rate of 28 samples per second. 2.4. Procedure The experimental sessions were conducted in a silent, dimly lit room. Prior to the experimental procedure, blood
2.5.2. Statistical analysis The task-induced changes in MCA flow velocities and their relationship with reaction time were evaluated using a three-way repeated measures ANOVA. Prior to this analysis, the sample was split according to the reaction time median of the sample (294.8 ms). The resulting performance groups (i.e. high vs. low reaction time) entered the model as between-subjects factor. Time (i.e. the 6 response intervals) and hemisphere (i.e. left MCA vs. right MCA) were included as within-subject factors. The values of dBFV for the 6 response intervals served as a dependent variable. The performance groups did not differ with regard to age (t = 0.56, p = .58), gender (U = 252, p = .31), BMI (t = 1.47, p = .15) and blood pressure (systolic: t = 1.41, p = .17; diastolic: t = 1.31, p = .20). This also held true for resting BFV in the left (t = 1.43, p = .16) and right
(t = 0.58, p = .56) MCAs. The mean reaction time was 262.9 ms (SD = 27.9) in the high and 346.9 ms (SD = 32.3) in the low performance group (t = 9.65, p < .001). The low performance group comprised 7 men and 17 women, while the high performance group comprised 4 men and 20 women. No significant differences between men and women were found with respect to reaction time (men: M = 288.3 ms, SD = 57.6 ms; women: M = 309.9, SD = 49.8, t = 1.22, p = .23) and the dBFV values for each of the 6 response intervals (all p > .10). Post hoc t-tests were applied to determine more precisely in which of the response intervals flow velocities in the MCA exceeded baseline level. For this purpose, single sample t-tests were used to test whether the dBFV values significantly differed from zero. In order to account for multiple testing (12 comparisons), alpha was set at .004. The 12 dBFV values were furthermore compared between both performance groups (alpha = .004). Finally, the values for the 6 response intervals were compared between both hemispheres (6 comparisons, alpha = .008). To complement the categorical analysis, stepwise multiple regression analyses were employed. The dBFV values for the 6 response intervals served as predictors and reaction time as dependant variable (alpha for t = .004). Two separate models were computed referring to the left and right MCAs. 3. Results Fig. 1 displays the time course of the BFV changes in the MCAs during the execution of the cued reaction time task. A steep bilateral increase of BFV was observed with a latency of less than 1 s after the cuing tone. A maximum appeared approximately 3.5 s after this stimulus followed by a continuous decline until the end of the registration period. The BFV increase was stronger in the right than in the left MCA beginning at second 2 after the cuing tone until the end of the registration period. In Fig. 2, the dBFV values for the 6 response intervals are shown for both performance groups. Except for the 1st second, the values were consistently higher in the high performance group. This was the case for both the left and right MCAs. The ANOVA revealed significant main effects for time (F[5, 230] = 64.99, p < .001; partial Eta squared = .59) and hemisphere (F[1, 46] = 7.82, p = .008; partial Eta squared = .15) which corroborated the overall taskinduced change in MCA flow velocities, as well as the right-hemispheric lateralization of the response. The factor performance group was also significant (F[1, 46] = 8.72, p = .005; partial Eta squared = .16) confirming the overall higher dBFV values in subjects with shorter reaction times. Furthermore, a significant interaction between the factors time and performance group was found (F[5, 230] = 3.57, p = .004; partial Eta squared = .072).
Change in flow velocity in %
S. Duschek et al. / Clinical Neurophysiology 119 (2008) 1292–1299
1295
Imperative stimulus
5 4 3 2
Cuing tone
Left MCA
1
Right MCA
0 -1
0
1
2
3
4
5
6
Time in s Fig. 1. Changes in MCA blood flow velocities during the execution of the cued reaction time task (grand average).
This indicated that the size of the group differences varied across the 6 time intervals. Time significantly interacted with hemisphere (F[5, 230] = 14.71, p < .001; partial Eta squared = .24) corroborating the stronger lateralization of the later as compared to the earlier phase of the hemodynamic response. The results of post hoc testing are given in Table 2. The dBFV values for the left and right MCAs differed significantly from zero (baseline level) in the case of each response interval except for the 1st second. In the seconds 2 and 3, significantly higher dBFV values for both vessels were found in the high performance group. In the right MCA, a significant group difference also emerged in second 4. The hemispheric difference was only significant in seconds 5 and 6. Table 3 displays the results of multiple regression analysis. Only the dBFV values for the third second significantly predicted reaction time. This held true for the models referring to both the left and the right MCAs. The dBFV values concerning the other response intervals were excluded in the first step, and no further models reached significance. 4. Discussion Among the brain areas relevant for the control of attentional arousal, the dorsolateral frontal and the inferior parietal lobes are parts of the perfusion territory of the MCA (Pardo et al., 1991; Paus et al., 1997; Haines, 2007). It seems plausible to ascribe the hemodynamic response observed during the execution of the reaction time task to neural activation processes in these structures. The more pronounced BFV increase in the right MCA emphasizes the dominance of the right hemisphere for arousal regulation (Posner and Petersen, 1990; Sturm et al., 1999).
1296
S. Duschek et al. / Clinical Neurophysiology 119 (2008) 1292–1299
Left MCA High performance Low performance
10
Change in flow velocity in %
Right MCA High performance Low performance
10
*
8
8
*
* 6
6
4
4
*
*
2
2
0
0
-2
-2 1
2
3
4
5
6
1
2
3
4
5
6
Time in s Fig. 2. Values of dBFV for both performance groups. Seconds 1–5 include the interstimulus interval. In second 6, the response was executed (bars represent standard deviations,* for p < .004).
The linkage between neural processes and cerebral hemodynamics is mediated by various physiological and biochemical mechanisms. At this, the tight coupling between nerve-cell activity and brain metabolism may play the most prominent role (Paulson, 2002; Iadecola, 2004). Controlled by a variety of neurochemical factors like K+, H+ and adenosine, neural activation leads to the dilation of cerebral arterioles and capillaries followed by an increase in local cerebral blood flow (Iadecola, 2004). Besides flow metabolism coupling, fast acting neural mechanisms contribute to cerebrovascular regulation. An intracranial vasodilative system is assumed which directly triggers the dilation of cortical microvessels as a response to brain stem activation (Sa´ndor, 1999; Sato et al., 2001). In animals, for instance, the experimental stimulation of
the Nucleus basalis Meynert of the brain stem was shown to produce cortical blood flow increase (Biesold et al., 1989). This mechanism is mediated by afferent fibres connecting lower brain areas with cortical microvessels (c.f. Sa´ndor, 1999; Hamel et al., 2002). In particular, cholinergic and serotonergic neurons are involved originating in the ascending reticular activating system (e.g. Sato et al., 2001; Szirmai et al., 2005). The ascending reticular activating system is of vast importance for the control of attentional arousal (Posner and Petersen, 1990), thus it is suggestive to assume an important role of the described neural system in the context of the present task. This rapid mechanism may be particularly involved in mediating the steep initial increase of BFV, which was observed with a latency of less than one
Table 2 Post hoc t-tests for differences of the dBFV values from zero (baseline), as well as for differences in the dBFV values between both performance groups and between both hemispheres Response interval
Baseline vs. task execution
High vs. low performance
*
*
t ( for p < .004)
t ( for p < .004)
Left vs. right MCA t (* for p < .008)
Second 1
Left MCA Right MCA
1.51 1.23
Left MCA Right MCA
0.58 0.66
0.57
Second 2
Left MCA Right MCA
3.03* 3.29*
Left MCA Right MCA
2.81* 2.81*
0.37
Second 3
Left MCA Right MCA
8.33* 9.28*
Left MCA Right MCA
3.15* 3.31*
1.72
Second 4
Left MCA Right MCA
8.49* 9.37*
Left MCA Right MCA
2.77 3.18*
2.34
Second 5
Left MCA Right MCA
7.48* 8.93*
Left MCA Right MCA
2.25 2.60
3.28*
Second 6
Left MCA Right MCA
5.99* 8.00*
Left MCA Right MCA
1.73 2.33
5.00*
S. Duschek et al. / Clinical Neurophysiology 119 (2008) 1292–1299 Table 3 Regression analyses for the prediction of reaction time from the dBFV values for the left and right MCAs, Beta and t-values (left MCA: R2 = .19, right MCA: R2 = .22) Response interval
Left MCA Beta
Second Second Second Second Second Second
1 2 3 4 5 6
.15 .12 .43 .05 .08 .13
*
t ( for p < .004) 1.03 0.48 3.23* 0.14 0.35 0.77
Right MCA Beta .16 .26 .47 .24 .02 .04
t (* for p < .004) 1.11 1.03 3.64* 0.71 0.07 0.24
second after the cuing signal. The contribution of both metabolic and non-metabolic factors to the hemodynamic response is also supported by the time-course of its lateralization across the task period. Given the higher metabolic activity in the right hemisphere related to preparatory attention (Sturm et al., 1999), flow metabolism coupling may be relevant mainly in the later phase of the response in which marked lateralization occurred. In contrast, non-metabolic factors may be dominant in triggering the non-lateralized initial response. Also an involvement of systemic blood pressure in the development of the cerebral hemodynamic response was reported (for overview see Duschek and Schandry, 2007). Autoregulatory processes are commonly assumed to keep cerebral blood flow constant by buffering transient blood pressure fluctuations (Chillon and Baumbach, 1997; Paulson, 2002). Nonetheless, the extent of cerebral hemodynamic modulation during mental activity has been shown to be moderated both by resting blood pressure and by cognitively induced blood pressure changes (Duschek and Schandry, 2004, 2006). Interestingly, the effects of blood pressure and heart rate on cerebral hemodynamic modulation during mental activity were also found to be most pronounced during the first seconds of the response (Duschek et al., 2008). In the present study, the task-induced increase in MCA flow velocities was substantially more pronounced in individuals who exhibited shorter reaction times. The interaction between the factors time and performance suggested differently pronounced effects of specific components of the hemodynamic response on reaction time. Post hoc testing based on an adjusted significance criterion indicated that the performance groups significantly differed with respect to the increase of BFV in the left MCA during seconds 2 and 3 after task onset. For the right MCA, a significant group difference was also found in second 4. Modulation of BFV in seconds 1, 5 and 6 did not differ significantly between groups. It is evident that the middle fraction of the hemodynamic response proved to predict reaction speed more precisely than did the later response component during which the motor reaction was executed. This is also supported by regression analysis which revealed that BFV modulation in both MCAs significantly predicted task performance only during the third second.
1297
The amounts of shared variance between BFV modulation and reaction time were 19% and 22% for the left and right vessels, respectively. This suggests a markedly closer relationship as compared to earlier fTCD studies using conventional analysis of hemodynamic modulation. Employing the same reaction time task, but averaging the BFV response in the MCAs across the entire interstimulus interval, yielded not more than 4% shared variance with reaction time (Duschek and Schandry, 2004). In a study based on a serial subtraction task, 9% common variance between overall task-induced increase in MCA flow velocities and arithmetic performance was found (Duschek and Schandry, 2006). The present finding is in line with the report of Schuepbach et al. (2007) on the relationship between BFV modulation and performance in mental planning. Schuepbach et al. (2007), however, found the closest association between MCA flow velocity modulation and task performance in an even earlier time window, i.e. the 2nd second after task onset. The present categorical analysis furthermore suggested a somewhat longer duration of the relevant time interval, namely 2–3 s. As an explanation of the divergent results, different time courses of the hemodynamic responses related to the varying cognitive requirements of the applied tasks may be considered. Schuepbach et al. (2007) observed a bilateral BFV maximum in second 2 of the 5-second interval during which mental planning took place. In the reaction time task, BFV peaked somewhat later, i.e. in second 4. One may speculate that this slower and longer lasting ascent of BFV was due to a continuous increase in attentional arousal during the preparation for the imperative stimulus. The higher magnitude of the later response component, in turn, could explain its closer relationship with performance. Nonetheless, the results of both studies are consistent in the sense that both clearly support the assumption of a specific and relatively early time window in which close relationships between BFV modulation and cognitive performance become apparent. The present study underlines that this notion is not restricted to executive functioning, but also applies to more basic cognitive processes such as preparatory attention. Also studies investigating cerebral blood flow using methods other than fTCD corroborated the specific relevance of the early hemodynamic response. For instance, volume-based hemodynamic measurements by means of optical imaging and laser Doppler flowmetry revealed a more precise spatial overlap with cortical field potentials during the first seconds of sensory stimulation than in later time intervals (Sheth et al., 2005). Functional brain imaging based on echoplanar MRI and gradient echo MRI suggested that ocular dominance columns may be most accurately resolved using the first few seconds of the BOLD response (Menon and Goodyear, 1999; Duong et al., 2000). Thus, also hemodynamic measurements based on blood volume and oxygenation underline the importance of the temporal dimension of cerebral blood flow modulation
1298
S. Duschek et al. / Clinical Neurophysiology 119 (2008) 1292–1299
for the investigation of brain behavior relationships (Sheth et al., 2005). One limitation of the study results from the restriction of the BFV recordings to the MCAs. Some of the brain areas relevant for the control of attentional arousal are not part of their perfusion territory. This holds particularly true for the anterior cingulate which is supplied by branches of the ACAs (Sturm et al., 1999; Haines, 2007). Activation processes in this structure are inevitably not reflected by the present measurement. In conclusion, the current study adds evidence to the notion of a time-locked relationship between rapid cerebral blood flow modulation and cognitive performance. In accordance with neurophysiological findings, the available fTCD studies suggest the existence of a time window of 2– 3 s after activation in which the cerebral hemodynamic response exhibits the closest linkage with neural activation, as well as basic and complex behavior. Future neurophysiological research in this area could certainly benefit from a stronger focus on the dynamic aspects of cerebral blood flow in addition to local hemodynamic distribution patterns. Acknowledgements We are grateful to Alessandro Angrilli and Ann-Kristin Adam for their help with the analysis of the hemodynamic data. References Ba¨cker M, Knecht S, Deppe M, Lohmann H, Ringelstein EB, Henningsen H. Cortical tuning: a function of anticipated stimulus intensity. Neuroreport 1999;10:293–6. Ba¨cker M, Deppe M, Zunker P, Henningsen H, Knecht S. Tuning to somatosensory stimuli during focal attention. Cerebrovasc Dis 1994;4(Suppl. 3):3. BeriSoft Cooperation. Experimental Run Time System, Version 3.32. Frankfurt a.M.: BeriSoft Cooperation, 2000. Biesold D, Inanami O, Sato A, Sato Y. Stimulation of the nucleus basalis of Meynert increases cerebral cortical blood flow in rats. Neurosci Lett 1989;98:39–44. Chillon JM, Baumbach GL. Autoregulation of cerebral blood flow. In: Welch KMA, Caplan LR, Reis DJ, Siesjo¨ BK, Weir B, editors. Primer on Cerebrovascular Diseases. San Diego: Academic Press; 1997. p. 51–4. Deppe M, Knecht S, Henningsen H, Ringelstein EB. AVERAGE: a Windows program for automated analysis of event related cerebral blood flow. J Neurosci Meth 1997;75:147–54. Deppe M, Ringelstein EB, Knecht S. The investigation of functional brain lateralization by transcranial Doppler sonography. Neuroimage 2004;21:1124–46. Duong TQ, Kim DS, Ugurbil K, Kim SG. Spatiotemporal dynamics of the BOLD fMRI signals: toward mapping submillimeter cortical columns using the early negative response. Magn Reson Med 2000;44:231–42. Duschek S, Schandry R. Functional transcranial Doppler sonography as a tool in psychophysiological research. Psychophysiology 2003;40:436–54. Duschek S, Schandry R. Cognitive performance and cerebral blood flow in essential hypotension. Psychophysiology 2004;41:905–13.
Duschek S, Schandry R. Deficient adjustment of cerebral blood flow to cognitive activity due to chronically low blood pressure. Biol Psychol 2006;72:311–7. Duschek S, Schandry R. Reduced brain perfusion and cognitive performance due to essential hypotension. Clin Auton Res 2007;17:69–76. Duschek S, Werner N, Kapan N, Reyes del Paso GA. Patterns of cerebral blood flow and systemic hemodynamics during arithmetic processing. J Psychophysiol 2008, in press. Frauenfelder BA, Schuepbach D, Baumgartner RW, Hell D. Specific alterations of cerebral hemodynamics during a planning task: a transcranial Doppler sonography study. Neuroimage 2004;22:1223–30. Haines DE. Neuroanatomy. An Atlas of Structures, Sections, and Systems. Philadelphia: Lippincott Williams & Wilkins; 2007. Hamel E, Vaucher E, Tong XK, St-Georges M. Neuronal messengers as mediators of microvascular toe in the cerebral cortex. ICS 2002;1235:262–76. Iadecola C. Neurovascular regulation in the normal brain and in Alzheimer’s disease. Nat Rev Neurosci 2004;5:347–60. Johnson A, Proctor RW. Attention, Theory and Practice. Thousand Oaks: Sage Publications; 2004. Knecht S, Henningsen H, Deppe M, Huber T, Ebner A, Ringelstein EB. Successive activation of both cerebral hemispheres during cued word generation. Neuroreport 1996;7:820–4. Logothetis NK, Pauls J, Trinath T, Oeltermann A. Neurophysiological investigation of the basis of the fMRI signal. Nature 2001;412:150–7. Menon RS, Goodyear BG. Submillimeter functional localization in human striate cortex using BOLD contrast at 4 Tesla: implications for the vascular point-spread function. Magn Resonan Med 1999;41:230–5. ¨ ber den Kreislauf des Blutes im menschlichen Gehirn. Leipzig: Mosso A. U Von Veit; 1881. Oldfield RC. The assessment and analysis of handedness: the Edinburgh inventory. Neuropsychologia 1971;9:97–113. Pardo JV, Fox PT, Raichle ME. Localisation of a human system for sustained attention by positron emission tomography. Nature 1991;349:61–4. Paulson OB. Blood–brain barrier, brain metabolism and cerebral blood flow. Eur Neuropsychopharm 2002;12:495–501. Paus T, Zatorre RJ, Hofle N, Caramanos Z, Gorman J, Petrides M, et al. Time-related changes in neural systems underlying attention and arousal during the performance of an auditory vigilance task. J Cogn Neurosci 1997;9:392–408. Posner MI, Rafal RD. Cognitive theories of attention and the rehabilitation of attentional deficits. In: Meier MJ, Benton AL, Diller L, editors. Neuropsychological Rehabilitation. Edinburgh: Churchill Livingston; 1987. p. 182–201. Posner MI, Petersen SE. The attention system of the human brain. Annu Rev Neurosci 1990;13:25–42. Roy C, Sherrington C. On the regulation of the blood supply of the brain. J Physiol 1890;11:85–108. Sa´ndor P. Nervous control of the cerebrovascular system: doubts and facts. Neurochem Int 1999;35:237–59. Sato A, Sato Y, Uchida S. Regulation of regional cerebral blood flow by cholinergic fibres originating in the basal forebrain. Int J Dev Neurosci 2001;19:327–37. Schuepbach D, Merlo MCG, Goenner F, Staikov I, Mattle HP, Dierks T, et al. Cerebral hemodynamic response induced by the tower of Hanoi puzzle and the Wisconsin card sorting test. Neuropsychologia 2002;40:39–53. Schuepbach D, Boeker H, Duschek S, Hell D. Rapid cerebral hemodynamic modulation during mental planning and movement execution: evidence of time-locked relationship with complex behavior. Clin Neurophysiol 2007;118:2254–62. Sheth SA, Nemoto M, Guiou MW, Walker MA, Toga AW. Spatiotemporal evolution of functional hemodynamic changes and their relationship to neuronal activity. J Cerebr Blood F Metab 2005;25:830–41.
S. Duschek et al. / Clinical Neurophysiology 119 (2008) 1292–1299 Stroobant N, Vingerhoets G. Transcranial Doppler ultrasonography monitoring of cerebral hemodynamics during performance of cognitive tasks: a review. Neuropsychol Rev 2000;10:213–31. Sturm W, de Simone A, Krause BJ, Specht K, Hesselmann V, Radermacher I, et al. Functional anatomy of intrinsic alertness: evidence for a fronto-parietal-thalamic-brainstem network in the right hemisphere. Neuropsychologia 1999;37:797–805.
1299
Szirmai I, Amrein I, Pa´lvo¨gyi L, Debreczeni R, Kamondi A. Correlation between blood flow velocity in the middle cerebral artery and EEG during cognitive effort. Cogn Brain Res 2005;24:33–40. Vingerhoets G, Stroobant N. Reliability and validity of day-to-day blood flow velocity reactivity in a single subject: an fTCD study. Ultrasound Med Biol 2002;28:197–202.