• Intraoperative optical imaging;
  • Human;
  • Seizures;
  • Perfusion;
  • Oximetry


  1. Top of page
  2. Methods
  3. Results
  4. Discussion
  5. Acknowledgments
  6. References

Purpose: Optical recording of intrinsic signals provides the highest combined spatial and temporal resolution with broad spatial sampling for measuring cerebral blood volume (CBV) and hemoglobin oxygenation in cerebral cortex. Few opportunities arise to apply this laboratory method to record spontaneous seizures in unanesthetized human brain during neurosurgery. We report such a rare opportunity in a man with recurrent focal epilepsy arising from a cavernous malformation.

Methods: We recorded intrinsic optical signals (IOS) from human cortex intraoperatively during spontaneous seizures arising from brain surrounding a small cavernous malformation in an awake patient using only local anesthesia with simultaneous electrocorticography. The IOS was recorded at two wavelengths, one an isosbestic point for hemoglobin to measure CBV (570 nm) and the other at a wavelength more sensitive to deoxygenated hemoglobin (Hbr) (610 nm). A modified Beer-Lambert calculation was used on two separate but similar seizures to approximate changes in Hbr, CBV as well as oxygenated hemoglobin (HbO2).

Results: Electrographically recorded seizures (n = 3) elicited a focal increase in both Hbr and CBV that lasted for the duration of the seizure, indicating that perfusion was inadequate to meet metabolic demand. Remarkably, these hemodynamic changes preceded the onset of the seizures by ∼20 s and occurred focally over the known location of the lesion and the seizure onsets.

Discussion: These findings demonstrate that the hemoglobin becomes deoxygenated in spite of large increase in CBV during spontaneous human focal seizures and that optically recorded hemodynamic events can be used both to predict and localize human focal epilepsy. Such data may someday be useful to assist in the presurgical evaluation of patients considered for epilepsy surgery and to predict the timing and location of seizure onsets.

Epilepsy is a disease of the brain characterized by recurrent spontaneous seizures. These “ictal” events arise from a population of hyperexcitable neurons that exhibit increased synchronized and desynchronized activity during the course of the seizure (Schwartzkroin, 1993). Certain aspects of the hemodynamic changes associated with focal seizures are well understood. The increase in neuronal activity leads to an increase the cerebral metabolic rate of oxygen (CMRO2) and glucose (CMRGluc), which causes an increase in cerebral blood flow (CBF) (Siesjo and Abdul-Rahman, 1979; Engel et al., 1982; Theodore et al., 1985; Ingvar, 1986; Van Paesschen, 2004). There is less consensus, however, on whether the increase in CBF is adequate to meet local metabolic demand and on the timing and focality of these hemodynamic events. Studies in animals and humans, using techniques with limited spatial and temporal resolution such as functional magnetic resonance imaging (fMRI), positron emission tomography (PET), single photon emission computer tomography (SPECT) and autoradiography generally show that the relative increase in CBF more than meets the increase in metabolism (Meldrum and Nilsson, 1976; Horton et al., 1980; Ingvar and Siesjo, 1983; Franck et al., 1986; Ingvar, 1986; Tanaka et al., 1990) leading to an increase in blood oxygenation (Blennow et al., 1979; Pinard et al., 1984; Jackson et al., 1994; Detre et al., 1995; Schwartz et al., 1998; Nersesyan et al., 2004). Studies using techniques with higher temporal resolution, such as oxygen-sensitive electrodes (Cooper et al., 1966; Dymond and Crandall, 1976), near infrared spectroscopy (NIRS) (Adelson et al., 1999) or optical recording of intrinsic signals (ORIS) (Suh et al., 2005a, 2005b; Bahar et al., 2006a, 2006b; Shariff et al., 2006), on the other hand, indicate that for an indefinite period of time CBF is inadequate to meet the metabolic demands of the epileptic neurons leading to a decrease in both tissue and hemoglobin oxygenation. However, significant controversy remains since cortical oxygenation may vary depending on systemic oxygenation, blood pressure, and seizure duration (Kreisman et al., 1983, 1984; Sokol et al., 2000). In addition, animal models of epilepsy generally involve acute pharmacologic alterations in the brain or electrically triggered seizures (Fisher, 1989) and studies are often performed under general anesthesia, which is known to effect cerebrovascular autoregulation (Sloan, 2002). Thus, whether focal seizures elicit a dip in tissue oxygenation in the unanesthetized human brain during chronic spontaneous seizures remains an open question. In addition, the focality and timing of the changes in perfusion and oxygenation are critical questions since these signals might be useful as surrogates for neuronal activity to map the seizure onset zone and possibly predict the onset of seizures for therapeutic intervention.

Our laboratory has developed the ability to perform ORIS intraoperatively on human brain during neurosurgical procedures (Schwartz et al., 2004; Schwartz, 2005; Suh et al., 2006). Although intraoperative ORIS currently offers the highest possible combined spatial (<200 μm) and temporal (33 ms) resolution of any human brain mapping technique, spontaneous seizures rarely occur intraoperatively. Fortuitously, we operated on a patient who had focal seizures that occurred with a predictable periodicity arising from face motor cortex and had the opportunity to perform ORIS during these seizures at multiple wavelengths in the absence of anesthesia. Our findings in this case confirm the animal data indicating that the increase in metabolic demand elicited by spontaneous focal human seizures is not adequately met by increases in CBF. In addition, we report that these focal alterations in cerebrovascular hemodynamics actually precede the onset of the seizures by ∼20 s, which may provide a novel mechanism for seizure prediction and a window for therapeutic intervention.


  1. Top of page
  2. Methods
  3. Results
  4. Discussion
  5. Acknowledgments
  6. References

The patient was a 45-year-old man with a 2 month history of progressive paroxysmal spasms of the left side of his face. Video-electroencephalographic monitoring revealed right frontal epileptiform discharges associated with the symptoms and an MRI scan demonstrated a small cavernous malformation in the face area of the right motor strip (Fig. 1A). Events occurred every 5 min, lasted for ∼80 s and were refractory to treatment with antiepileptic medications. The patient was brought to the operating room for a stereotactic awake craniotomy with motor mapping, electrocorticography (ECoG) and resection of the lesion and surrounding epileptogenic cortex. For ECoG, a 20-contact grid was placed over the known stereotactic location of the cavernous malformation and the surrounding gyri. Five seizures were recorded, all arising from the same two contacts overlying the gyrus containing the cavernous malformation, which coincided with face motor cortex. Once the mapping was completed and the location of the lesion and epileptic cortex identified, optical imaging of intrinsic signals (ORIS) was performed prior to resection in accordance with a protocol approved by the Institutional Review Board at Weill Medical College of Cornell University. Only local anesthetic was used during the imaging.


Figure 1. (A) Gradient echo axial MRI scan demonstrates a small cavernous malformation the right motor strip. (B) Surface of the brain under glass footplate. The black circles highlight the location of the recording electrodes. The rectangles demonstrate three regions of interest (ROIs), which contained the pixel values with the most statistically significant changes for each of the three seizures. The label on the rectangle corresponds to the graphs within this figure. (C) ECoG recording of a typical seizure. Scale bars: 20 seconds and 1 millivolt. The time course of (D and F) oximetry and (E) perfusion related intrinsic optical signal calculated as -ΔR/R (%) during each seizure from each ROI in (B) is graphed along with the power of the ECoG. Error bars represent SD of pixel values from each ROI. The onset of statistically significant optical signal changes is indicated with a black arrow and the onset of significant change in the power of the ECoG is indicated with a gray arrow. (G, H, and I). ORIS maps for each of the three seizures in (D), (E) and (F) at the time of the maximum change in the intrinsic signal. (J and K) Hbr (J) and Hbt (K) maps calculated using modified Beer-Lambert law on seizures shown in (D) and (E). Note the location of the maximal optical change is similar to the raw 610 nm (G) and 570 nm (H) images. Scale bar: 1 cm.

Download figure to PowerPoint

Electrocorticography recording

A 4-contact grid electrode (interelectrode distance, 1 cm; Ad-Tech, Madison, WI, U.S.A.) was placed on surface of the cortex, partially overlapping the gyrus containing the cavernous malformation and known region of seizure onset (Fig. 1B). ECoG data was amplified (Grass IP511) and digitized at 2,000 Hz (CED Power 1401, Cambridge, UK), and recorded onto a computer using Spike 2 (Cambridge Electronic Design, Cambridge, U.K.). The electrocardiogram was similarly digitized and recorded.

Optical imaging procedure

A sterile glass footplate (4 cm × 4 cm) was placed over the grid and adjacent cortex from which the seizures arose to dampen the movement artifacts caused by heartbeat and respiration. Care was taken not to blanch the vessels from the pressure of the footplate (Fig. 1B). A custom-made camera holder with gross and fine x-y-z- manipulators was used to suspend the camera over the surface of the brain (Imager 3001, analog camera, Optical Imaging Inc., Germantown, NY, U.S.A.). A single 35 mm lens was used for image acquisition since a back-to-back lens system did not provide an adequate field of view (Ratzlaff, 1991; Schwartz et al., 2004). We performed 3 × 3 binning of the 768 × 480 pixel resolution to reduce the size of each dataset. The operating room was darkened and the cortex illuminated with a ring illuminator attached to Tungsten halogen lamp (100 W). The light was passed through bandpass filters of different wavelengths, first at 570 ± 10 nm to record the surface blood vessel pattern and then at 610 ± 10 nm and 570 ± 10 nm for ORIS. The optical reflectance signal was digitized onto a computer at 33 frame/s.

Since the patient was having spontaneous seizures every 5 min and each seizure lasted <2 min, we designed the trials to begin data acquisition prior to the onset of the seizure and last for 5 min to ensure that the entire duration of the seizure would be recorded (Fig. 1C). Since the seizures were spontaneous, we could not control the exact time of onset of each seizure within our trial. One entire seizure was recorded at a single wavelength and then the following seizure was recorded at the other wavelength in an alternating pattern.

We chose to record at 570 and 610 nm based on their sensitivity to Hbt and Hbr, respectively. Since HbO2 and Hbr absorb light equally at the isosbestic wavelength of 570 nm, the change in reflection is proportional to total hemoglobin (Hbt) or CBV (Sheth et al., 2003) and CBF, assuming that the concentration of red blood cells remains constant (Vanzetta and Grinvald, 2001; Nemoto et al., 2004). At 610 nm, the majority of the signal arises from the oxygenation state of hemoglobin, since Hbr absorbs light with three times the absorption coefficient of HbO2 (Mayhew et al., 2000).

Although the IOS recorded at 610 nm is more sensitive to Hbr, they are not equivalent, since changes in HbO2 may influence this relationship. The actual concentration of Hbr, Hbt, and HbO2 can be calculated if data is acquired simultaneously at both 570 and 610 nm using a modified form of the Beer-Lambert calculation (Sheth et al., 2003; Suh et al, 2006). In this report we were only able to acquire data from different wavelengths sequentially during separate seizures. Although in a prior report of human IOS we have demonstrated that following cortical stimulation the signal recorded at 610 nm is highly correlated with the corrected Hbr signal, it is not clear if this holds true of epileptiform events (Suh et al., 2006). Therefore, in order to best approximate the actual changes in concentration in Hbr, Hbt, and HbO2 we applied the modified form of the Beer-Lambert calculation to two sequential seizures of similar length, one acquired at 610 nm and the other at 570 nm after time-locking the onset of each seizures.

Optical signal data analysis

The change in reflectance during each seizure was calculated by dividing each frame during each trial by a baseline, which consisted of an average of 300 frames (150 frames [4.5 s] from the onset of each trial and 150 frames at the end of each trial). The baseline frames in each trial occurred at least 20 s before the onset of the seizure and 1 min after the offset of the seizure. We chose the baseline both at the beginning and end of the trial to minimize any slow changes in the hemodynamic signal that might occur over the 5 min of each trial such as vasomotor noise. Changes in reflection of light were calculated as −ΔR/R (%) from a region of interest (ROI) that contains the pixel values with the most statistically significant changes compared with baseline. (Matlab). Since only local anesthetic was used, seizures were often accompanied by head motion. This mechanical noise was eliminated using a semiautomatic computer program. Using the relative positions of the recording electrodes, blood vessels, and other cortical surface details, we generated a baseline image to which all other frames were compared. Each frame was corrected by a combination of whole-pixel shifts to account for integer offsets to the surface features, and Fourier-shifts1 to account for fractional offsets. The mean pixel difference (MPD) was calculated for each frame, and any frame whose MPD was more than 1.5 times greater than the temporal median was subjected to a third motion correction. Each of these frames was Fourier-shifted using 20 intervals linearly spaced between −2.5% and 2.5% of the image's total pixel height and width, providing 400 shifted frames for comparison to the baseline frame. Two-dimensional correlation coefficients were calculated for each frame, the maximum coefficient was noted, and each frame was Fourier-shifted by the appropriate amount. The end result of these motion correction steps is to minimize high-frequency fluctuations in the pixel values of the intrinsic signal caused by cranial motion rather than hemodynamic changes. However, it is possible that changes in the intrinsic signal were artificially minimized as a result of the algorithm. This is of particular concern in the cases where two-dimensional correlation coefficients were used—the shifted frame with the highest correlation is the best mathematical fit, although not necessarily the best structural fit. So, while the signal-to-noise ratio is certainly increased, the signal is also potentially decreased. The area of activation was calculated from pixels with a change in reflection 3 SD above or below the mean during baseline recording using a parametric t-test and a p-value <0.05.

Electrocorticography data analysis

The time of seizure onset was determined from the ECoG, which was bandpass (3–70 Hz) filtered. The power of ECoG was calculated and the onset and offset of the seizure were identified based on statistically significant change (p < 0.05) from baseline (mean and SD during first and last five seconds of the recording) using a two-tailed t-test.


  1. Top of page
  2. Methods
  3. Results
  4. Discussion
  5. Acknowledgments
  6. References

Three spontaneous seizures were successfully recorded, two at 610 nm and one at 570 nm. Seizures were detected every 5 min and lasted 78.3 ± 5.5 s (n = 3) (Fig. 1C). Each seizure was accompanied by a dramatic, focal change in the intrinsic signal.

At 610 nm, a significant increase light reflectance began 23.74 ± 8.67 s prior to the electrographic onset of the seizure (Figs. 1 and 2). The signal was confined to the gyrus from which the seizures arose. The signal maximum reached 14.2 ± 0.4% and peaked 38.4 ± 34.8 s after the onset of the seizures. The change in reflectance lasted for ∼200 s beyond the termination of the seizure. This increase in reflectance most likely represents an increase in Hbr consistent with a “dip” in hemoglobin oxygenation (see below). The maximal spatial extent of the signals, thresholded at 3 SD below the baseline, was 148.9 ± 73.9 mm2. The spatial maps of the two seizures recorded at 610 nm were remarkably similar, centered over the crest of the gyrus containing the cavernous malformation with an area of 148.9 mm2, thresholded at 3 SD above the baseline (Figs 1G and 1I).


Figure 2. Changes in optical signal and ECoG during 40 seconds before and after the onset of seizures. (A, B and C) Optical signals from a linear ROI (60 × 6 pixels (2.3 × 0.23 cm)) along the crest of the gyrus are graphed above. Spectral analysis of ECoG (middle graph) and raw ECoG (bottom) are displayed for each of the three seizures. For the spectral analsysis graph, the frequency component of the fourier transform of each ECoG tracing was color-coded and plotted as a function of time. For the second 610 nm seizure (B), the image acquisition before the onset of seizure was shorter than 40s, since the seizure onsets were spontaneous, hence the grey bar represents the absence of optical data.

Download figure to PowerPoint

At 570 nm, a significant (p < 0.05) decrease in light reflectance began 15.0 s prior to the electrographic onset of the seizure (Figs. 1 and 2), again restricted to the known epileptic gyrus consistent with a focal drop in CBV. Prior to the onset of the seizures, the signal inverted to a significant increase in light reflectance (increase in CBV) which reached a maximum amplitude of 46.2%, peaking 58.1 s after the onset of the seizure. The increase in CBV persisted ∼180 s after the offset of the seizure. The area of maximal signal change was still on the same gyrus, but slightly anterior with an area of 94.2 mm2, which when compared with the Hbr signal was 36% smaller in size. These results indicate that a decrease in CBV precedes the onset of spontaneous seizures and that CBV, although slower to develop, is larger in amplitude and equally as good as Hbr at localizing the region of seizure onset.

In order to estimate the precise change in concentration of Hbr, HbO2 and Hbt, a modified Beer-Lambert law was applied to two subsequent seizures recorded at 610 and 570 nm wavelengths. After the pathlength correction at each wavelength, Hbr maximum reached 12.7% at 35.3 s and Hbt maximum reached 13.8% at 62.4 s after the onset of the seizures. These values are similar to the values obtained using raw images at 610 and 570 nm (see above; Figs. 1J and 1K). Also, as shown in raw images obtained at 610 nm, Hbr increase persisted well beyond the termination of the seizure. For the better comparison of perfusion and oximetry signal changes in raw images versus corrected images, ORIS movies are shown in APPENDIX.

In order to examine the relationship between the ECoG and the optical signal, we graphed the power of each frequency in the ECoG for a period of 40 s prior to and 40 s after the onset of the seizure compared with a linear ROI of the pixel changes during this same time period to see if there were any early pre-ictal changes in the ECoG which might explain the anticipatory optical signal (Fig. 2). We also looked statistically at eight 5-s bins of ECoG data from 40-0 s prior to the onset of the seizures to see if there were any changes in the power of the ECoG prior to the onset of each seizure (ANOVA, p > 0.05). These results indicate that the early optical signal changes were not a response to an electrographic change.


  1. Top of page
  2. Methods
  3. Results
  4. Discussion
  5. Acknowledgments
  6. References

In this study we show that focal changes in perfusion and oxygenation precede the onset of spontaneous human seizures by ∼20 s and may be useful at predicting the onset and location of seizures prior to any electrographic change. A simultaneous decrease in hemoglobin oxygenation and a decrease in CBV occur within the gyrus of known ictal onset. At the electrographic and clinical onset of the seizure, the decrease in hemoglobin oxygenation becomes more profound, remaining restricted to the epileptic gyrus, but accompanied by a focal increase in CBV, also restricted to the epileptic gyrus. These changes persist after the offset of the seizure for at least 1–2 min. These results are extremely significant, since the intraoperative occurrence of spontaneous seizures is quite rare during optical recording sessions (Haglund and Hochman, 2004). For this reason, the few prior reports have been limited to interictal events or electrically triggered afterdischarges (Haglund et al., 1992; Haglund and Hochman, 2005). Similarly, ictal studies of hemoglobin oxygenation using fMRI in humans are also limited by the infrequency of spontaneous ictal events and associated movement artifact, manifested by the plethora of papers on interictal events and rarity of ictal reports (Jackson et al., 1994; Detre et al., 1995; Schwartz et al., 1998; Salek-Haddadi et al., 2002).

Whether CBF is adequate to meet the increased metabolic demands of a seizure has been a long-standing debate. The initial hypoxia-hypoperfusion hypothesis, based on histologic similarities between ischemic and epileptic brain damage, hypothesized that the cell damage following status epilepticus (SE) was caused by cerebral anoxia (Norman, 1964; Plum et al., 1968; Meldrum et al., 1973; Meldrum and Nilsson, 1976; Soderfeldt et al., 1983; Simon, 1985). Later studies refuted this theory based on findings that the percent increase in CBF was greater than the increase in cerebral metabolism (Horton et al., 1980; Ingvar and Siesjo, 1983; Ingvar, 1986; Tanaka et al., 1990), the cellular damage associated with SE was not identical to hypoxic injury (Siesjo and Abdul-Rahman, 1979; Siesjo, 1981; Siesjo and Wieloch, 1986), ictal increases in venous oxygenation (Plum et al., 1968; Caspers and Speckmann, 1972; Vern et al., 1976; Pinard et al., 1984), oxidation in the mitochondrial transport chain, NADH and cytochrome oxidase (Jobsis et al., 1971; Mayevsky and Chance, 1975; Vern et al., 1976; Hempel et al., 1980), increases in tissue pO2 (Kreisman et al., 1981a, 1981b, 1983a, 1983b, 1984) and that tissue injury occurred even in the absence of cerebral anoxia (Meldrum and Nilsson, 1976; Pinard et al., 1984; Meldrum, 2002). Likewise, human studies with ictal PET, SPECT, and fMRI demonstrate increases in CBF beyond that exceed metabolic demand (Kuhl et al., 1980; Franck et al., 1986) as well as increases in blood oxygenation levels dependent (BOLD) signals (Jackson et al., 1994; Detre et al., 1995; Schwartz et al., 1998; Salek-Haddadi et al., 2002). However, these studies were performed using techniques with inferior temporal or spatial resolution and limited spatial sampling.

Multiwavelength ORIS can measure blood oxygenation and CBV from large areas of cortex simultaneously with a spatial and temporal resolution limited by the specifications of the camera acquiring the data and the evolution of the signal (33 ms and <200 μm). Our data indicates that during spontaneous focal seizures in the unanesthetized human brain, focal decreases in hemoglobin oxygenation persist throughout the duration of the seizure in spite of a large increase in CBV. In the operating room, oxygen is being administered though a nasal cannula, and both blood pressure and systemic oxygen are continuously monitored and maintained stable. Thus, systemic hypoxia or hypotension cannot be responsible. Although we did not directly measure metabolism or tissue pO2, we presume that the increase in CMRO2 associated with the seizure surpasses the compensatory vasodilatory mechanisms of the brain. It is unlikely that the cavernous malformation altered local autoregulatory mechanisms since it is a low flow vascular anomaly. Although an increase in Hbr does not mandate a decrease in tissue pO2, our results are consistent with prior measurements of tissue oxygenation in unanesthetized humans with implanted oxygen-sensitive electrodes, which have demonstrated decreases in tissue pO2 during complex partial seizures (Cooper, 1963, 1966; Dymond and Crandall, 1976). Likewise, animal data using ORIS and oxygen-sensitive electrodes in a model of neocortical focal epilepsy corroborates these findings, demonstrating an increase in Hbr and a decrease in tissue pO2 for variable lengths of time at the onset of each seizure (Suh et al., 2005; Bahar et al., 2006a, 2006b; Shariff et al., 2006). It is curious that investigators using interictal and ictal spike-triggered fMRI generally report an increase in the BOLD signal, which would be consistent with a decrease in Hbr, i.e. the opposite of our finding (Krakow et al., 2001; Benar et al., 2002; Jager et al., 2002; Salek-Haddadi et al., 2002; Aghakhani et al., 2003; Salek-Haddadi et al., 2006). We attribute this inconsistency not only to the lower temporal and spatial resolution of fMRI but the fact that most fMRI studies have been performed on interictal rather than ictal events which may elicit such a brief focal increase in Hbr as to be undetectable without higher strength magnets (Krakow et al., 2001; Benar et al., 2002; Jager et al., 2002; Salek-Haddadi et al., 2006). Likewise, prior ictal fMRI studies have been done on generalized spike-and-wave events, which may not elicit an increase in Hbr in the cortex (Salek-Haddadi et al., 2002; Aghakhani et al., 2003) or are performed without concurrent electrical recordings rendering the timing unclear (Jackson et al., 1994; Detre et al., 1995; Schwartz et al., 1998).

For all three seizures, the optical data localized the onset and spread of the seizure to the appropriate gyrus. Although the oxygen-related signal peaked earlier, the amplitude of the perfusion-related signal was much larger, indicating a higher signal-to-noise ratio. There has been debate in the literature regarding which signal is best for localizing epileptiform events (Haglund and Hochman, 2004; Suh et al., 2005, 2006; Bahar et al., 2006a, 2006b) and our data demonstrate that, at least for lesional neocortical epilepsy, either signal can be used successfully. Currently, the gold standard in mapping ictal events in preparation for surgery is with the chronic implantation of subdural arrays of electrodes (Lüders et al., 1992). Electrical recordings from the brain surface can be inaccurate based on volume conduction and inaccurate modeling assumptions (Buzsáki and Traub, 1997). Chronic optical mapping of light absorption may provide complementary or even superior localizing information. Miniaturized light emitting and recording equipment currently exists to explore this question.

Several novel therapeutic methodologies such as cortical stimulation (Kossoff et al., 2004), focal drug perfusion (Eder et al., 1997) and cooling (Rothman et al., 2005) are currently under investigation to terminate seizures. All of these therapies would be more efficacious if a “closed-loop” system provided ongoing feedback of cortical physiology and could reliably predict the onset of seizures. Complex mathematical algorithms applied to electrographic data for seizure prediction have recently garnered much attention (Litt and Echauz, 2002). The hemodynamic events we describe in this case report may provide an alternative method. The notion that hemodynamic events may be useful at predicting the onset of seizures has been proposed as early as 1933 (Gibbs, 1933). More recent studies using transcranial Doppler have demonstrated increases in lobar perfusion as early as 20 min before focal as well as generalized spike-and-wave events (Weinand et al., 1994; De Simone et al., 1998; Diehl et al., 1998). Likewise, both increases (Adelson et al., 1999; Salek-Haddadi et al., 2002; Makiranta et al., 2005) and decreases (Hoshi and Tamura, 1992) in tissue oxygenation have been found tens of seconds before seizure onset using fMRI and NIRS. Our data confirms and extends these hypotheses with extremely high temporal and spatial resolution measurements demonstrating statistically significant focal preictal alterations in cerebral hemodynamics in the region of seizure onset. These pre-ictal changes include a decrease and then increase in CBV and a dip in hemoglobin oxygenation. We have not found similar results in our animal experiments in anesthetized models of acute pharmacologically induced neocortical epilepsy (Suh et al., 2005a, 2005b; Bahar et al., 2006a, 2006b; Shariff et al., 2006), which may be an artifact of the model, species or anesthesia. In contrast, the data we present here represents spontaneous seizures from unanesthetized human cortex and are the most applicable to the clinical situation.

The etiology of preictal hemodynamic changes are unknown. One possibility is that the vascular changes are not preictal but rather ictal and the ECoG recording is not sufficiently sensitive to record subtle electrical events. A single electrode not only suffers from sampling limitations, only measuring from a small radius around the electrode, but also volume conduction (Buzsáki and Traub, 1997; Lauritzen, 2001; Logothetis et al., 2001; Schwartz and Bonhoeffer, 2001; Sheth et al., 2003; Suh et al., 2005). Early events such as “fast ripples” although brief, may be missed by surface electrodes (Bragin et al., 2002). In addition, ictal changes in the ECoG are a reflection of synchronous dendritic activity in a large population of neurons. On the other hand, the hemodynamic changes recorded with the IOS arise mostly from subthreshold activity in an area 5–10 times larger than the area of spiking cells (Grinvald et al., 1994; Das and Gilbert, 1995; Toth et al., 1996; Bosking et al., 1997). Thus early subtle ictal activity may not be recorded by the ECoG electrode but clearly alter the IOS. Another possibility is that the preictal changes are not electric events in neurons but rather caused by other changes in the local environment such as extracellular fluid shifts, astrocyte swelling or alterations in the concentration of other mediators of vasodilation such as lactate, carbon dioxide, nitrous oxide or potassium which might change light reflection or hemodynamics without markedly effecting the ECoG signal (Kuschinsky, 1978; Lauritzen, 2001; Erinjeri and Woolsey, 2002; Iadecola, 2004). Other possibilities might include astrocyte-mediated vasodilatation spread via calcium waves (Takano et al., 2006).

The main limitation of our study is that it is a report of a single case. However, the results were reproducible over three seizures in this individual. The methodology of acute human intraoperative ORIS has been validated by our laboratory as well as others (Sato et al., 2002; Pouratian et al., 2003; Haglund and Hochman, 2004; Suh et al., 2006). With our set-up we have demonstrated reproducible maps of oxygenation and blood flow following a controlled reproducible stimulus, namely focal cortical bipolar electrical stimulation (Suh et al., 2006). Another limitation is that we were unable to record multiple wavelengths simultaneously and thus could not quantify the increases in Hbr and Hbt. Roughly, 10–30% of the signal recorded at 610 nm comes from increases in CBV and may be artificially increased by an enormous vasodilation (Sheth et al., 2004). For this reason we estimated the actual changes in Hbr and Hbt by using the modified Beer-Lambert calculation on sequential similar seizures recorded at 570 or 610 nm to show that the raw data is not significantly different than the modified data, as we had shown in previous human studies (Suh et al., 2006).

In summary, we have shown that the local increase in metabolic demand during spontaneous focal neocortical seizures surpasses the compensatory local increase in CBF causing the deoxygenation of hemoglobin. These hemodynamic changes, although small, are focal and precede the onset of seizures by ∼20 s offering clinical promise in both seizure localization and prediction.

  • 1

    Given an image f(x, y) and its two-dimensional Fourier-transformed counterpart F(μ, ν), then the shifted image f(x-α, y-β) is equivalent to F(μ, ν) ei (αμ+βν).


  1. Top of page
  2. Methods
  3. Results
  4. Discussion
  5. Acknowledgments
  6. References

We thank Jonathan Victor for editorial assistance with the manuscript.


  1. Top of page
  2. Methods
  3. Results
  4. Discussion
  5. Acknowledgments
  6. References
  • Adelson PD, Nemoto E, Scheuer M, Painter M, Morgan J, Yonas H. (1999) Noninvasive continuous monitoring of cerebral oxygenation periictally using near infrared spectroscopy: a preliminary report. Epilepsia 40:14841489.
  • Aghakhani Y, Bagshaw AP, Benar CG, Hawco C, Andermann F, Dubeau F, Gotman J. (2003) fMRI activation during spike and wave discharges in diopathic generalized epilepsy. Brain 127:11271144.
  • Bahar S, Suh M, Zhao M, Schwartz TH. (2006a) Intrinsic optical signal imaging of neocortical seizures: The “epileptic dip.” Neuroreport 17:499503.
  • Bahar S, Suh M, Mehta AD, Schwartz TH. (2006b) In vivo intrinsic optical signal imaging of neocortical epilepsy. In BroderickPA, RahniDN, KolodnyEH. (Eds) Bioimaging in neurodegeneration. Humana Press, Totowa , NJ , pp. 149175.
  • Benar C-G, Goross DW, Wang Y, Petre V, Pike B, Dubeau F, Gotman J. (2002) The BOLD response to interictal epileptiform discharges. NeuroImage 17:11821192.
  • Blennow G, Folbergrova J, Nilsson B, Siesjo BK. (1979) Cerebral metabolic and circulatory changes in the rat during sustained seizures induced by DL-homocysteine. Brain Res 179:129146.
  • Bosking WH, Zhang Y, Schofield B, Fitzpatrick D. (1997) Orientation selectivity and the arrangement of horizontal connections in tree shrew striate cortex. J Neurosci 17:21122127.
  • Bragin A, Mody I, Wilson CL, Engel JJ. (2002) Local generation of fast ripples in epileptic brain. J Neurosci 22:20122021.
  • Buzsáki G, Traub RD. (1997) Physiological basis of EEG activity. In EngelTA, PedleyTA. (Eds) Epilepsy: a comprehensive textbook. Lippincott-Raven, Philadelphia , pp. 819830.
  • Caspers H, Speckmann EJ. (1972) Cerebral pO2, pCO2 and pH changes during convulsive activity and their significance for spontaneous arrest of seizures. Epilepsia 13:699725.
  • Cooper R. (1963) Local changes in of intracerebral blood flow and oxygen in humans. Med Biol Eng 1:529536.
  • Cooper R, Crow HJ, Walter WG, Winter AL. (1966) Regional control of cerebral vascular reactivity and oxygen supply in man. Brain Res 3:174191.
  • Das A, Gilbert CD. (1995) Long-range horizontal connections and their role in cortical reorganization revealed by optical recording of cat primary visual cortex. Nature 375:780784.
  • De Simone R, Silvestrini M, Marciani MG, Curatolo P. (1998) Changes in cerebral blood flow velocities during childhood absence seizures. Pediatr Neurol 18:132135.
  • Detre JA, Sirven JI, Alsop DC, O'Connor MJ, French JA. (1995) Localization of subclinical ictal activity by functional magnetic resonance imaging: correlation with invasive monitoring. Ann Neurol 38:618624.
  • Diehl B, Knecht S, Deppe M, Young C, Stodieck SR. (1998) Cerebral hemodynamic response to generalized spike-wave discharges. Epilepsia 39:12841289.
  • Dymond AM, Crandall PH. (1976) Oxygen availability and blood flow in the temporal lobes during spontaneous epileptic seizures in man. Brain Res 102:191196.
  • Eder HG, Stein A, Fisher RS. (1997) Interictal and ictal activity in the rat cobalt/pilocarpine model of epilepsy decreased by local perfusion of diazepam. Epilepsy Res 29:1724.
  • Engel JJ, Kuhl DE, Phelps ME. (1982) Patterns of human local cerebral glucose metabolism during epileptic seizures. Science 218:6466.
  • Erinjeri JP, Woolsey TA. (2002) Spatial integration of vascular changes with neural activity in mouse cortex. J Cereb Blood Flow Metab 22:353360.
  • Fisher RS. (1989) Animal models of the epilepsies. Brain Res Rev 14:245278.
  • Franck G, Sadzot B, Salmon E, Depresseux JC, Grisar T, Peters JM, Guillaume M, Quaglia L, Delfiore G, Lamotte D. (1986) Regional cerebral blood flow and metabolic rates in human focal epilepsy and status epilepticus. Adv Neurol 44:935948.
  • Gibbs FA. (1933) Cerebral blood flow preceding and accompanying experimental convulsions. Arch Neurol Psychiat Chicago 30:10031010.
  • Grinvald A, Lieke EE, Frostig RD, Hildesheim R. (1994) Cortical point-spread function and long-range lateral interactions revealed by real-time optical imaging of macaque monkey primary visual cortex. J Neurosci 14:25452568.
  • Haglund MM, Hochman DW. (2004) Optical imaging of epileptiform activity in human neocortex. Epilepsia 45:4347.
  • Haglund MM, Hochman DW. (2005) Furosemide and mannitol suppression of epileptic activity in the human brain. J Neurophysiol 94:907918.
  • Haglund MM, Ojemann GA, Hochman DW. (1992) Optical imaging of epileptiform and functional activity in human cerebral cortex. Nature 358:668671.
  • Hempel FG, Kariman K, Saltzman HA. (1980) Redox transition in mitochondia of cat cerebral cortex with seizures and hemorrhagic hypotension. Am J Physiol 238:H249H258.
  • Horton RW, Meldrum BS, Pedley TA, McWilliam JR. (1980) Regional cerebral blood flow in the rat during prolonged seizure activity. Brain Res 192:399412.
  • Hoshi Y, Tamura M. (1992) Cerebral oxygenation state in chemically-induced seizures in the rat—study by near infrared spectrophotometry. Adv Exp Med Biol 316:137142.
  • Iadecola C. (2004) Neurovascular regulation in the normal brain and in Alzheimer's disease. Nat Rev Neurosci 5:347360.
  • Ingvar M. (1986) Cerebral blood flow and metabolic rate during seizures. Relationship to epileptic brain damage. Ann NY Acad Sci 462:207223.
  • Ingvar M, Siesjo BK. (1983) Local blood flow and glucose consumption in the rat brain during sustained bicuculline-induced seizures. Acta Neurol Scand 68:129144.
  • Jackson GD, Connelly A, Cross JH, Gordon I, Gadian DG. (1994) Functional magnetic resonance imaging of focal seizures. Neurology 44:850856.
  • Jager L, Werhahn KJ, Hoffmann A, Berthold S, Scholz V, Weber J, Noachtar S, Reiser M. (2002) Focal epileptiform activity in the brain: detection with spike-related functional MR Imaging—preliminary results. Radiology 223:860869.
  • Jobsis FF, O'Connor M, Vitale A, Vreman H. (1971) Intracellular redox changes in functioning cerebral cortex. I. Metabolic efects of epileptiform activity. J Neurophysiol 34:735741.
  • Kossoff EH, Ritzl EK, Politsky JM, Murro AM, Smith JR, Duckrow RB, Spencer DD, Bergey GK. (2004) Effect of an external responsive neurostimulator on seizures and electrographic discharges during subdural electrode monitoring. Epilepsia 45:15601567.
  • Krakow K, Messina D, Lemieux L, Duncan JS, Fish DR. (2001) Functional MRI activation of individual interictal epileptiform spikes. NeuroImage 13:502505.
  • Kreisman NR, LaManna JC, Rosenthal M, Sick TJ. (1981a) Oxidative metabolic responses with recurrent seizures in rat cerebral cortex: role of systemic factors. Brain Res 218:175188.
  • Kreisman NR, Sick TJ, LaManna JC, Rosenthal M. (1981b) Local tissue oxygen tension—cytochrome aa3 redox relationships in rat cerebral cortex in vivo. Brain Res 218:161174.
  • Kreisman NR, Rosenthal M, LaManna JC, Sick TJ. (1983a) Cerebral oxygenation during recurrent seizures. Adv Neurol 34:231239.
  • Kreisman NR, Sick TJ, Rosenthal M. (1983b) Importance of vascular responses in determining cortical oxygenation during recurrent paroxysmal events of varying duration and frequency of repetition. J Cereb Blood Flow Metab 3:330338.
  • Kreisman NR, Sick TJ, Rosenthal M. (1984) Concepts of brain oxygen sufficiency during seizures. Adv Exp Med Biol 180:381392.
  • Kuhl DE, Engel J Jr, Phelps ME, Selin C. (1980) Epileptic patterns of local cerebral metabolism and perfusion in humans determined by emission computed tomography of 18FDG and 13NH3. Ann Neurol 8:348360.
  • Kuschinsky WWM. (1978) Local chemical and neurogenic regulation of cerebral vascular resistance. Physiol Rev 58:656689.
  • Lauritzen M. (2001) Relationship of spikes, synaptic activity, and local changes of cerebral blood flow. J Cereb Blood Flow Metab 21:13671383.
  • Litt B, Echauz J. (2002) Prediction of epileptic seizures. Lancet Neurol 1:2230.
  • Logothetis NK, Pauls J, Augath M, Trinath T, Oeltermann A. (2001) Neurophysiological investigation of the basis of the MRI signal. Nature 412:150157.
  • Lüders H, Awad I, Burgess R, Wyllie E, Van Ness P. (1992) Subdural electrodes in the presurgical evaluation for surgery of epilepsy. Epilepsy Res 5(Suppl):147156.
  • Makiranta M, Ruohonen J, Suominen K, Niinimaki J, Sonkajarvi E, Kiviniemi V, Seppanen T, Alahuhta S, Jantti V, Tervonen O. (2005) BOLD signal increase preceeds EEG spike activity—a dynamic penicillin induced focal epilepsy in deep anesthesia. NeuroImage 27:715724.
  • Mayevsky A, Chance B. (1975) Metabolic responses of the awake cerebral cortex to anoxia, hypoxia, spreading depression and epileptiform activity. Brain Res 98:149165.
  • Mayhew JEW, Johnston D, Berwixk J, Jones M, Cofey P, Zheng Y. (2000) Spectroscopic analysis of neural activity in brain: increased oxygen consumption following activation of barrel cortex. NeuroImage 12:664675.
  • Meldrum BS. (2002) Concept of activity-induced cell death in epilepsy: historical and contemporary perspectives. Prog Brain Res 135:311.
  • Meldrum BS, Nilsson B. (1976) Cerebral blood flow and metabolic rate early and late in prolonged epileptic seizures induced in rats by bicuculline. Brain 99:523542.
  • Meldrum BS, Vigouroux RA, Brierley JB. (1973) Systemic factors and epileptic brain damage. Prolonged seizures in paralyzed, artificially ventilated baboons. Arch Neurol 29:8287.
  • Nemoto M, Sheth S, Guiou M, Pouratian N, Chen JWY, Toga AW. (2004) Functional signal- and paradigm-dependent linear relationships between synaptic activity and hemodynamic responses in rat somatosensory cortex. J Neurosci 24:38503861.
  • Nersesyan H, Hydeer F, Rothman DL, Blumenfeld H. (2004) Dynamic fMRI and EEG recording during Spike-Wave seizures and generalized tonic-clonic seizures in WAG/Rij rats. J Cereb Blood Flow Metab 24:589599.
  • Norman RM. (1964) The neuropathology of status epilepticus. Med Sci Law 4:4651.
  • Pinard E, Tremblay E, Ben-Ari Y, Seylaz J. (1984) Blood flow compensates oxygen demand in the vulnerable CA3 region of the hippocampus during kainate-induced seizures. Neuroscience 13:10391049.
  • Plum F, Posner JB, Troy B. (1968) Cerebral metabolic and circulatory responses to induced convusions in animals. Arch Neurol 18:113.
  • Pouratian N, Sheth SA, Martin NA, Toga AW. (2003) Shedding light on brain mapping: advances in human optical imaging. Trends Neurosci 26:277282.
  • Ratzlaff EH, Grinvald A. (1991) A tandem-lens ep: fluorescence macroscope: hundred-fold brightness advantage for wide-field imaging. J Neurosci Methods 36:127137.
  • Rothman SM, Smyth MD, Yang XF, Peterson GP. (2005) Focal cooling for epilepsy: an alternative therapy that might actually work. Epilepsy Behav 7:214221.
  • Salek-Haddadi A, Merschhemke M, Lemieux L, Fish DR. (2002) Simultaneous EEG-Correlated Ictal fMRI. NeuroImage 16:3240.
  • Salek-Haddadi A, Diehl B, Hamandi K, Merschhemke M, Liston A, Friston K, Duncan JS, Fish DR, Lemieux L. (2006) Hemodynamic correlates of epileptiform discharges: an EEG-fMRI study of 63 patients with focal epilepsy. Brain Res 1088:148–166.
  • Sato K, Nariai T, Sasaki S, Yazawa I, Mochida H, Miyakawa N, Momose-Sato Y, Kamino KYO, Hirakawa K, Ohno K. (2002) Intraoperative intrinsic signal imaging of neuronal activity from subdivisions of the human primary somatosensory cortex. Cerebral Cortex 12:269280.
  • Schwartz TH, Bonhoeffer T. (2001) In vivo optical mapping of epileptic foci and surround inhibition in ferret cerebral cortex. Nature Med 7:10631067.
  • Schwartz TH. (2005) The application of optical recording of intrinsic signal to simultaneously acquire functional, pathological and localizing information and its potential role in neurosurgery. Stereotac Funct Neurosurg 83:3644.
  • Schwartz TH, Resor SR Jr, De La Paz R, Goodman RR. (1998) Functional magnetic resonance imaging localization of ictal onset to a dysplastic cleft with simultaneous sensorimotor mapping: intraoperative electrophysiological confirmation and postoperative follow-up: technical note. Neurosurgery 43:639644; discussion 644–635.
  • Schwartz TH, Chen LM, Friedman RM, Spencer DD, Roe AW. (2004) Intraoperative optical imaging of human face cortical topography: a case study. NeuroReport 15:15271531.
  • Schwartzkroin PA. (1993) Epilepsy. Models, mechanisms and concepts. Cambridge University Press, Cambridge .
  • Shariff S, Suh M, Zhao M, Ma H, Schwartz TH. (2006) Recent developments in oximetry and perfusion-based mapping techniques and their role in the surgical treatment of neocortical epilepsy. Epilepsy Behav 8:363375.
  • Sheth SA, Nemoto M, Guiuo M, Walker M, Pouratian N, Toga AW. (2003) Evaluation of coupling between optical intrinsic signals and neuronal activity in rat somatosensory cortex. NeuroImage 19:884894.
  • Sheth SA, Nemoto M, Guiou G, Walker M, Pouratian N, Toga AW. (2004) Linear and nonlinear relationships between neuronal activity, oxygen metabolism, and hemodynamic response. Neuron 42:347355.
  • Siesjo BK. (1981) Cell damage in the brain: a speculative synthesis. J Cereb Blood Flow Metab 1:155185.
  • Siesjo BK, Abdul-Rahman A. (1979) A metabolic basis for the selective vulnerability of neurons in status epilepticus. Acta Physiol Scand 106:377378.
  • Siesjo BK, Wieloch T. (1986) Epileptic brain damage: pathophysiology and neurochemical pathology. Adv Neurol 44:813847.
  • Simon RP. (1985) Physiologic consequences of status epilepticus. Epilepsia 26(suppl 1):S58S66.
  • Sloan TB. (2002) Anesthetics and the brain. Anesthesiol Clin North Am 20:265292.
  • Soderfeldt B, Blennow G, Kalimo H, Olsson Y, Siesjo BK. (1983) Influence of systemic factors on experimental epileptic brain injury. Structural changes accompanying bicuculline-induced seizures in rats following manipulations of tissue oxygenation or alpha-tocopherol levels. Acta Neuropathol (Berl) 60:8191.
  • Sokol DK, Markand ON, Daly EC, Luerssen TG, Malkoff MD. (2000) Near infrared spectroscopy (NIRS) distinguishes seizure types. Seizure 9:323327.
  • Suh M, Bahar S, Mehta AD, Schwartz TH. (2005a) Temporal dependence in uncoupling of blood volume and oxygenation during interictal epileptiform events in rat neocortex. J Neurosci 25:6877.
  • Suh M, Shariff S, Bahar S, Mehta AD, Schwartz TH. (2005b) Intrinsic optical signal imaging of normal and abnormal physiology in animals and humans—seeing the invisible. Clin Neurosurg 52:135149.
  • Suh M, Bahar S, Mehta AD, Schwartz TH. (2006) Blood volume and hemoglobin oxygenation response following electrical stimulation of human cortex. NeuroImage 31:6675.
  • Takano T, Tian G-F, Peng W, Lou N, Libionka W, Han X, Nedergaard M. (2006) Astrocyte-mediated control of cerebral blood flow. Nat Neurosci 9:260267.
  • Tanaka S, Sako K, Tanaka T, Nishihara I, Yonemasu Y. (1990) Uncoupling of local blood flow and metabolism in the hippocampal CA3 kainic acid-induced limbic seizure status. Neuroscience 36:339348.
  • Theodore WH, Brooks R, Margolin R, Patronas N, Sato S, Porter RJ, Mansi L, Bairamian D, DiChiro G. (1985) Positron emission tomography in generalized seizures. Neurology 35:684690.
  • Toth LJ, Rao SC, Kim DS, Sur M. (1996) Subthreshold facilitation and suppression in primary visual cortex revealed by intrinsic signal imaging. Proc Natl Acad Sci USA 93:98699874.
  • Van Paesschen W. (2004) Ictal SPECT. Epilepsia 45:3540.
  • Vanzetta I, Grinvald A. (2001) Evidence and lack of evidence for the initial dip in the anesthetized rat: implications for human functional brain imaging. NeuroImage 13:959967.
  • Vern B, Schuette WH, Whitehouse WC, Matsuga N. (1976) Cortical oxygen consumption and NADH fluoresceince during metrazol seizures in normotensive and hypotensive cats. Exp Neurol 52:8398.
  • Weinand ME, Carter LP, Patton DD, Oommen KJ, Labiner DM, Talwar D. (1994) Long-term surface cortical cerebral blood flow monitoring in temporal lobe epilepsy. Neurosurgery 35:657664.