Both Middle Cerebral Artery (MCA) mean flow velocity measured by transcranial color Doppler (TCCD) and somatosensory evoked potential (SSEP) show high interhemispheric symmetry in normal subjects [1-4]. This condition could be modified after subarachnoid haemorrhage (SAH), due to prevalent unilateral alterations in vasoregulation of cerebral arteries and vasospasm [5, 6], resulting in impaired regional brain perfusion [7, 8]. The vessel section πr² (r = radius of the vessel) is inversely proportional to mean blood flow velocity V, and directly proportional to blood flow F (πr²= F/V). The variability in MCA vessel section resistances and blood flow, between hemispheres within one subject, can be evaluated by a semi quantitative approach, comparing the reciprocal of MCA mean blood flow velocities (1/vMCA) detected simultaneously on both sides, assuming the other determinants of F and V to be constant (F = ΔP∙πr²∙r²/8ηl, V = ΔP∙πr²/8ηl, ΔP = cerebral perfusion pressure, η = blood viscosity and l = length of measured segments of MCA). In addiction, there is strong correlation between variations of electrophysiological changes and regional cerebral blood flow (rCBF), and a 50% reduction in cortical SSEP amplitude at rCBF levels between 14 and 16 ml/100 g/min has been demonstrated [9]; for this reason cortical N20 SSEP from the median nerve has been widely used to evaluate cerebral function in carotid surgery and intracranial aneurysm operative procedures [10-12]. Mechanism, diagnosis and prevention of brain injury after SAH have been widely investigated [13-20], but evaluation of correlation between variations of cortical activity expressed by evoked potentials and regional brain vascular changes detected by transcranial color-Doppler is lacking. The aim of the study is to investigate the correlation between changes in resistances due to vessel section reduction of MCA and modifications of cortical SSEP by interhemispheric comparison of both values of 1/vMCA detected by TCCD and N20 amplitude, in a series of case studies of patients with SAH.

During the survey all 16 patients were eligible. A total of 66 bilateral SSEP and TCCD were performed with the criteria of simultaneousness. Values of the 1/v MCA and N20 amplitude recorded on the compromised hemisphere of each patient, expressed as ratios of their respective contralaterals and compared to CT/MRI scan images of the corresponding MCA supplied area, N20 amplitude absolute value, presence/absence of vasospasm and lateralizing symptoms are shown in Tab.3. The 1/vMCA and N20 interhemispheric ratio of 40 examinations were analyzed in respect of 40 images of CT/MRI scan performed on the same day as the TCCD and SSEP. Sixteen examinations in eleven patients (group 1: patients No. 1,3,5,6,9,10,12,13,14,15,16) matched CT/MRI image of no brain damage in the area supplied by the MCA: all subjects showed a high level of symmetry between interhemispheric values of 1/vMCA and N20 amplitude (Fig. 1); both interside ratios were always >0,65 and showed correlation (p< 0,01, r = 078, 95% CI = 0,46 to 0,92). In this group, MCA unilateral vasospasm (mean flow velocity > 120 < 150 cm /sec, L.I.>3) or almost symmetric bilateral vasospasm (< 180 cm/sec, L.I.>3) was detected in 18,7% of examinations; N20 amplitudes were still bilaterally normal (>1,2 μV) and symmetric. Patients did not show lateralizing neurological defects or consciousness alteration at clinical evaluation (Table 4). Fourteen examinations in seven patients (group 2: patients No. 2,3,5,7,12,15,16) matched CT/MRI images of tissue alterations in unilateral MCA area (shift, oedema, tissue hypodensity, cloth) or hydrocephalus. Both values of 1/vMCA and N20 amplitude on the compromised side were lower than contralateral and both interhemi-spheric ratios were always <0,65 in all these patients (p <0,01 r = 0,67, 95% CI = 0,23 to 0,88) (Fig. 2). Unilateral MCA vasospasm (mean flow velocity > 120 < 180 cm/sec and L.I. >3) was detected in 57% of overall examinations. N20 amplitudes on the compromised hemisphere were much lower than those on the contralateral in all patients, but still > 1,2 μV, except in patient No.2, who reached pathological amplitude (<1,2 μV) in several examinations. Corresponding lateralizing neurological defect of muscular strength or conciousness alteration was detected at clinical examination in all these patients (Table 5). Ten examinations in four patients (group 3: patients No.2,4,8,11) matched unilateral CT/MRI brain structural ischemic or hemorrhagic damage evidence; in all of them, resistance (1/vMCA) on the compromised side reverted to a high ratio to the contralateral (>0,65) and no longer correlated to the persistent interhemispheric asymmetry of N20 amplitude (ratio < 0,65) (Fig. 3); in this group correlation between the interhemispheric ratios of 1/MCA and N20 was lost (p=0.88, r = -0.05, 95% CI = -0.66 to 0.59). MCA vasospasm (mean flow velocity > 120< 165 cm /sec) was detected in 33,3% of examinations. The N20 value on the compromised hemispheres was pathologic (<1,2 μV) or absent in 9 examinations and all patients had stable hemiparesis or hemiplegia (Table 6). Patients No. 2, 3,5,12,15 belong to more than one group at different times, as a consequence of modification of their brain CT/MRI images during the survey; patient No.2 shifted from group 2 to 3; patients No. 3 and No.5 shifted from group 2 to 1; patients No.12 and No.15 shifted from group 1 to 2 and back to 1 again. In addition, we examined the course and correlation between the two interhemispheric ratios in those patients, seven in all, who were submitted to more than three examinations on different days. Statistic correlation (p<0.01) emerged in patient No. 9 (Fig. 4), who had no CT modifications and in those who showed transient unilateral alterations on CT/MRI in MCA supplied brain area and resumed a normal picture (patients No. 3, 5,12,15; Figs. 5-8). Interhemispheric modifications of 1/vMCA and N20 amplitude did not correlate in patients who had morphological brain damage on CT scan immediately after SAH (patient No.11, p = 0.89) (Fig. 9) or who subsequently developed brain injury (patient No.2, p = 0.2) (Fig. 10a,b).

Fig. (1). The couples of N20 and 1/vMCA interhemispheric ratio values, detected in each patient without CT/MRI scan alterations in the MCA supplied area, gathered in Table 4, are displayed in the graph in descending order. All N20 and 1/vMCA interhemispheric ratio values were >0.65 and showed significant statistical correlation ( p < 0.01, r = 0.78, 95% C.I. = 0.46 to 0.92).

Fig. (2). The couples of N20 and 1/vMCA interhemispheric ratio values, detected in each patient with CT/MRI scan unilateral transient alterations in the MCA supplied area, gathered in table 5, are displayed in the graph in descending order. All N20 and 1/vMCA interhemispheric ratio values were < 0.65 and showed significant statistic correlation (p < 0.01, r = 0.67, 95% C.I. = 0.23 to 0.88).

Fig. (3). The couples of N20 and 1/vMCA interhemispheric ratio values detected in each patient with CT/MRI scan unilateral structural ischemic or hemorrhagic alterations in the MCA supplied area, gathered in table 6, are displayed in the graph in descending order of N20 ratio values . The two ratios show divergent courses, since the value of N20 on the compromised hemisphere is < 0.65 with respect to the contralateral, whereas the 1/vMCA value reverts to ratio >0.65. The two ratios no longer correlate in this group ( p = 0.88, r = -0.05, 95% C.I.= - 0.66 to 0.59).

Fig. (4). Patient No. 9 showed parallel courses of N20 and 1/vMCA interhemispheric ratios ( always > 0.65), with significant statistical correlation ( p <0.01, r = 0.94, C.I. 0.57 to 0.99). This patient never developed CT/MRI scan alteration of MCA area. He showed transient vasospasm, even bilateral, (BFV >120 < 240), normal N20 amplitude and was always asymptomatic during the survey.

Fig. (5). Patient No. 3 showed parallel courses of N20 and 1/vMCA interhemispheric ratios with significant statistical correlation ( p < 0.01, r =0.91, C.I. 0.64 to 0.98). The patient showed TC scan image of parietal oedema corresponding to both ratio values <0.65 in the first examination (point 1); this promptly recovered to ratio values >0.65 after oedema disappeared (point 2). This patient never reached the vasospasm threshold, always maintained N20 amplitude > 1,2 µV on the compromised hemisphere and showed clinical evidence of lateralizing strength deficit with respect to the first examination only (point 1).

Fig. (6). Patient No. 5: the courses of N20 and 1/vMCA ratios correlate (p < 0.01, r = 0.94, C.I. 0.65 to 0.99). He showed both ratios < 0.65 correspondig to points 1 to 6 (a) and and CT/MRI images of parietal oedema or hypodensity (b, points 3 and 4); the last couple of values > 0.65 (a, point 7) correspond with a re-established normal CT scan (b, point 7). He had vasospasm (BFV >120 <180) corresponding to points 2,3 and 4; he always maintained N20 absolute value > 1,2 µV on the compromised hemisphere; when both ratios reached a value < 0.65 (points 2 to 6), he showed lateralizing symptoms and/or consciousness modification; when both ratios reverted to a value > 0.65 (point 7) he recovered to a normal clinical condition. (N20 amplitude is expressed in µV and MCA blood flow velocity in cm/sec).

Fig. (7). Patient No. 12 showed parallel courses of N20 and 1/vMCA ratios (p <0.01, r = 0.92, C.I. = 0.44 to 0.99). He showed hydrocephalus promptly treated with EVD (point 1) and right parietal oedema (point 2) in MCA area at CT scan (both ratios < 0.65). He maintained N20 amplitude > 1.2 µV on the compromised hemisphere and vasospasm (BFV >120 to 180 cm/sec) was detected corresponding to points 2,3,4,5; the patient had clinical transient lateralizing symptoms corresponding to point 2 and recovered a normal clinical picture in subsequent examinations (points 3 to 6); CT scan performed at point 6 was normal (both ratios > 0.65).

Fig. (8). Patient No. 15: the parallel variations of N20 and 1/vMCA ratio (a) show a statistically significant correlation (p < 0.01, r = 0.94, C.I. 0.58 to 0.99); at first examination both ratio values > 0.65 (a, point 1) corresponded with normal MCA supplied brain area at CT scan (b, point 1); later modifications of both ratios to values < 0.65 (a, points 2,3) corresponded with the appearance of an image of parietal hypodensity at CT scan (b, point 3); the following ratio values > 0.65 corresponded to normalized CT scan (b, point 6). This patient had vasospasm at point 2 (BFV >180<200 cm/sec), had constant > 1.2 µV N20 amplitude, and showed lateralizing symptoms and/or consciousness modification at points 2 and 3 (ratios < 0.65). (N20 amplitude is expressed in µV and MCA blood flow velocity in cm/sec).

Fig. (9). This patient showed left parietal hemorrhagic damage evident at CT scan immediately after SAH: the courses of N20 and 1/vMCA ratios were divergent and did not correlate (p = 0.89), since N20 had pathologic amplitude < 1.2 µV at point 1 and 2 and disappeared in subsequent examinations, whereas 1/vMCA was always almost symmetric on hemispheres. The patient never had vasospasm and clinical stable hemiparesis was evident.

Fig. (10). Patient No. 2 showed parallel courses of very low N20 and 1/v MCA ratio (<0.65) detected in the first three examinations (a, points 1to 3), corresponding with CT image of hypodesity of the right capsula (b, point 2); in the subsequent examinations, the divergent courses of N20 (< 0.65) and 1/vMCA (>0.65) and non correlated ratios (p= 0.2 r = 0.46, C.I. for r : -0.28 to 0.86) corresponded with CT and MRI images of structural infarct in the same area (b, point 4 and 8). Right N20 amplitude was < 1.2µV at points 1to 3 and 6 to7 (see examinations 2,3,4,7,8 in Table 3) and =< 1.5 µV at points 4,5, 8and 9 (see examinations 5,6,9,10 in Table 3); right MCA vasospasm (BFV >120 <180 cm/sec) was detected at points 1to 6. Clinical hemiparesis was always evident. (N20 amplitude is expressed in µV and MCA blood flow velocity in cm/sec).

The coupling between neuronal activity and its blood supply is of critical importance to the physiology of the human brain. Subarachnoid hemorrhage causes high rates of neurological deficits following impaired regional brain perfusion (15), mainly due to alterations of vascular resistances and pressure autoregulation. The vessel section, πr², is inversely proportional to mean blood flow velocity V, and directly proportional to blood flow F (πr²= F/V). In simultaneous detection of bilateral MCA mean blood flow velocities, within one subject, we can assume constant the CPP, blood viscosity and length of measured segments of MCA. Consequently, changes in vascular resistance and blood flow, between hemispheres, due to vessels section reduction of MCA, can be evaluated in an indirect way, by a semi quantitative approach, comparing 1/vMCA detected simultaneously on both sides. Likewise, interhemispheric differences in cortical activity of the primary somatosensory area can be evaluated comparing the amplitude of the N20 waves. Our data demonstrate close correlation between the variations of 1/vMCA and cortical N20 amplitude, matching their respective interhemispheric ratios. Modifications of interhemispheric ratio of 1/vMCA correspond with analogue variations of interhemispheric ratio of N20 amplitude in a wide range of values, from 0.9 to 0.4-0.3 (p< 0.01), unless structural ischemic or haemorrhagic damage develops (Fig. 11). When 1/vMCA and N20 amplitude had symmetric interhemispheric values (both ratios >0.65) CTI/MRI scan showed absence of alterations in the MCA areas. When both values of 1/vMCA and N20 amplitude on the compromised hemisphere became <0.65 of the contralateral, CT/MRI scan images showed transient modifications in the affected MCA supplied brain area and clinical evaluation revealed corresponding lateralizing symptoms or consciousness alteration (patients No. 2,3,5,12,15); consequently this data seems to identify the ischemic risk threshold. This condition of ischemic penumbra is potentially reversible, until variations of 1/MCA and N20 amplitude correlate, even at interhemispheric ratio values of 0.3- 0.4, if such extreme variations prevail only for a short time. In fact, when both 1/vMCA and N20 amplitude were observed to have regained ratio >0.65 to contralateral values (patients No. 3, 5, 12, 15), the CT/MRI scan images of MCA area and clinical condition returned to normal. On the other hand, if both low ratio values persist, as shown in patient 2, who reached interhemispheric values <0.4 for three consecutive days, structural damage develops. When evident CT/MRI structural brain damage occurs subsequently (patient No.2), or immediately after SAH (patients No. 4, 8, 11), there is a lack of correlation, as pathologic amplitude (<1, 2 μV) or absence of N20, on the compromised side, reflects the tissue damage and determined stable interhemispheric ratio <0.65. Conversely, in this situation 1/vMCA reverted to a ratio >0.65 with respect to the contralateral. This lack of correlation is probably due to the loss of neurovascular coupling and suggests a luxury perfusion in ischemic areas which no longer vasoregulate. The three patterns described above are shown in the scatter diagram (Fig. 12). A previous study investigated correlation between SSEP Central Conduction Time (CCT) and vasospasm and found a statistically significant (p < 0.01) increase in actual CCT (6.7 msec) in only the severe grade of vasospasm (blood flow velocity >200 cm/s) in subarachnoid hemorrhage [24]. By contrast, in our approach, variations of 1/vMCA and N20 cortical amplitude showed correlation before damage developed and independently from vasospasm. The high correspondence between variation in vascular resistances due to vessel section and N20 amplitude can be explained by the close correlation between blood flow and electric activity in the cortical areas. In fact, the cortex, whose activity is expressed by the N20 amplitude, has greater ischemic sensitivity than the thalamus and medial lemniscus, which are involved in CCT [25]. In an experimental study of graded cerebral ischemia on an animal model, Burnett and colleagues demonstrated that the reduction in SSEP amplitude corresponds to a progressive regional CBF (rCBF) decrease and even a small reduction of as little as 10% in rCBF produces analogue modification in SSEP [26]. Similarly, in healthy subjects, linear coupling and linear covariation between N20-P22 amplitude and intensity of blood oxygen level-dependent functional magnetic resonance imaging (BOLD- fMRI) has been demonstrated [27, 28]. Correspondence between variations of 1/vMCA and N20 cortical amplitude in our data emerges whether vasospasm is detected or not, as it was detected in 23% of examinations of group 1 and in 57% of group 2. Recent studies reviewed by Pluta and colleagues suggest that vasospasm of larger arteries is only to be an epiphenomenon of early brain injury after SAH with consequent microvascular vasospasm, spreading depression and cortical dysfunction [8] and the difficulty in defining the correlation between vasospasm and ischemia has been highlighted: perfusion abnormality without macrovascular vasospasm in the watershed areas or in the vicinity of sulcal clots has been demonstrated [29]. Dankbaar and colleagues reported that vasospasm decreases cerebral perfusion, but corresponds with the least perfused region in only two thirds of patients; furthermore, almost half of patients with severe vasospasm do not have delayed cerebral ischemia (DCI), and, although severe vasospasm can decrease perfusion, it may not result in DCI [30]. Carrera and colleagues found that 40% of patients who developed DCI did not reach blood flow velocity (BFV) >120 cm/sec and 16% of patients never registered BFV >120 [31]. Some limitations of our study have to be considered. First, we considered a small number of patients and examinations (SSEP,TCCD and neuroimaging) were timed on the clinical basis evaluation, not following a specific protocol. Second, the semi quantitative comparison of data from both hemispheres can be missed if data change to the same pathologic extent bilaterally. On the basis of our approach, symmetric bilateral pathological variation of vascular resistances of MCA would correspond to analogue symmetric variation of N20 amplitude: in this case we could not evaluate the ischemic risk by only comparing the two ratios. We can hypothesize that a very low bilateral N20 amplitude, close to the pathological threshold of 1,2 μV, could be detected; however we found no such condition of symmetric bilateral pathological changes in our case series. Third, we evaluated modifications of MCA brain supplied area by CT or MRI, who detect changes with a different sensitivity; moreover, recently Dankbaar and colleagues demonstrated that CT perfusion (CTP) is more sensitive in detecting ischemic penumbra than CT [32], so its use could help to better define the three interhemispheric patterns of neurovascular coupling.

Fig. (11). Couples of N20 and 1/vMCA ratio values detected in patients of group 1 and group 2, who had no CT/MRI image of structural ischemic or haemorrhagic injury in MCA supplied area, displayed in descending order. Modifications of interhemispheric ratio of 1/vMCA correspond with analogue variations in interhemispheric ratio of N20 amplitude in a wide range of values, from 0.9 to 0.4-0.3, (p<0.01, r = 0.93, Cl for r = 0.89 to 0.96 unless structural ischemic or haemorrhagic damage develops.

Fig. (12). The scatter diagram shows the distribution of all 66 couples of 1/vMCA and N20 ratio values gathered in Table 3. Each point identifies a couple of 1/vMCA (X axis) and N20 (Y axis) interhemispheric ratio values. In section A, both 1/vMCA and N20 interhemispheric ratio values > 0.65 correspond to CT/MRI normal images of MCA supplied area: this is the area of absence of damage, where N20 amplitude and 1//vMCA values show interhemispheric symmetry. In section B, 1/vMCA ratios > 0.65 coupled with N20 ratios < 0.65 correspond with CT/MRI images of unilateral structural brain damage in the MCA supplied area: this is the brain injury area, where pathological amplitude or absence of N20 reflects the tissue damage, whereas 1/vMCA values reverting to those of the uninjured contralateral area suggests a luxury perfusion state. In section C, both 1/vMCA and N20 interhemispheric ratio values < 0.65 correspond with CT/MRI images of unilateral transient and reversible alterations in MCA supplied area; this condition is capable of reverting to normal (section A) or evolving to stable damage (section B): this section could be named “ ischemic penumbra zone”. In section D no ratio allocation is found: this area would correspond with a condition of unilateral hemispheric alteration of vascular resistances due to reduction of MCA vessel section and consequent low 1/v MCA ratio, in contrast to symmetrical bilateral N20 amplitude and high interhemispheric ratio value. It is reasonable to hypothesize an impossible coexistence of altered regional brain perfusion with normal cortical activity.

MCA vascular changes detected by TCCD in SAH show significant statistic linear correlation with cortical activity measured by the early cortical component (N20) of the somatosensory evoked potential from the median nerve. Three different patterns of neurovascular coupling were identified matching the interhemispheric ratios of N20 and reciprocal of MCA mean flow velocity detected within each patient, to CT scan or MRI images of MCA areas. 1) 1/vMCA and N20 ratios > 0.65 correspond with absence of neuroimaging alterations in MCA areas and lateralizing symptoms: this pattern indicates an ongoing equilibrium of neurophysiological and vascular condition. 2): both 1/vMCA and N20 values lower than 0.65 in respect of the contralaterals correspond with TC/MRI scan images of transient alterations in the MCA area and clinical evidence of lateralizing symptoms or consciousness modification: this pattern identifies a situation of reversible ischemic penumbra and can evolve either to the re-establishment of a normal and symmetric condition, expressed in pattern 1, where N20 and 1/vMCA absolute values show little or no interhemispheric difference; or to the next pattern. 3) 1/vMCA value > 0.65 and N20 amplitude < 0.65 in respect of the contralaterals correspond with TC/MRI scan images of hemispheric structural damage in the MCA area: this is a condition of evident structural ischemic or hemorrhagic damage in which there is a luxury perfusion in the injured brain area. Although more studies are required to confirm our hypothesis, we would like to emphasize the importance of simultaneous evaluation of somatosensory evoked potential and the TCCD to monitor the neurovascular coupling and to detect the ischemic penumbra due to MCA vascular changes in SAH, before the ischemic damage develops.

Written consent was obtained from the patients for publication of these case series and any accompanying images. A copy of the written consent is available for review from the Editor-in-Chief of this journal.

PDP conceived the work, collected and analysed the data as well as collaborated in writing the article. PZ analyzed the data and collaborated in writing the article. MI analyzed the data and collaborated in writing the article. EB analyzed the data. EF analyzed the data and elaborated the statistical calculations. All authors have read and approved the final manuscript.