Utility of Quantitative EEG in Neurological Emergencies and ICU Clinical Practice

Abstract

The electroencephalogram (EEG) is a cornerstone tool for the diagnosis, management, and prognosis of selected patient populations. EEGs offer significant advantages such as high temporal resolution, real-time cortical function assessment, and bedside usability. The quantitative EEG (qEEG) added the possibility of long recordings being processed in a compressive manner, making EEG revision more efficient for experienced users, and more friendly for new ones. Recent advancements in commercially available software, such as Persyst, have significantly expanded and facilitated the use of qEEGs, marking the beginning of a new era in its application. As a result, there has been a notable increase in the practical, real-world utilization of qEEGs in recent years. This paper aims to provide an overview of the current applications of qEEGs in daily neurological emergencies and ICU practice, and some elementary principles of qEEGs using Persyst software in clinical settings. This article illustrates basic qEEG patterns encountered in critical care and adopts the new terminology proposed for spectrogram reporting.

Keywords: cyclic patterns; frequency domain; qEEG; rhythmic patterns; seizures; spectral analysis; spectrogram; time domain.

Figure 1

(A) qEEG amplitude measures. For all qEEG measures, the left hemisphere is depicted in blue and the right hemisphere in red; time is represented on the x-axis and amplitude in different scales on the y-axis. 1. Amplitude integrated EEG (aEEG): the maximum and minimum amplitude of each epoch connected with a vertical line is depicted using a semi-logarithmic scale (linear from 0 to 10 µV, and logarithmic from 10 to 100 μV). 2. Envelope trend: the peak amplitude from each epoch is plotted. 3. Suppression percentage: shows the percentage of suppression ranging from 0% (no suppression) to 100% (complete suppression). At the beginning, the raw EEG recording (B) and all amplitude EEG trends show normal amplitudes, and progressively the amplitude decreases as is depicted by the drop on the aEEG, peak envelope, and rise in suppression percentage. At the end of the recording, there is a complete suppression of the EEG (C).

Figure 2

(A) An EEG shown in a classical EEG visualization; the time domain: x-axis 1s/division scale and y-axis 70 μV/division. (B) Fourier transform of a segment of an EEG (shadow rectangle on (A)). (C) Spectrogram: a color scale is applied to the Fourier transform (from white and warm colors for higher power to cooler colors (black) for lower power) and time is represented on the x-axis in minutes or even hours, with frequency on the y-axis and color on the z-axis. Each epoch of Fourier analysis becomes a column of pixels. In this normal EEG recording, the posterior dominant alpha rhythm produces a peak of power at 10 Hz in the Fourier transform; this is clearly visible on the color code as a red–yellow line at 10z in the spectrogram, indicating the maximal power at that frequency.

Figure 3

(A) An EEG shown in the time domain. (B) Fourier transform of a segment of an EEG (shadow rectangle on (A)) with a delta brush; clearly recognizable is a main peak of power at delta frequencies that arrives at red colors in the color scale, and other peak at beta frequencies that arrive at green colors in the color scale. (C) Spectrogram with a clear increase of power in the delta band (red band) and a less pronounced increase of power (green band) in beta frequencies that conform to the brush. It is interesting to note that if the frequency analysis had been cut at 20 Hz, the peak of the beta component (27 Hz, in this specific case) would have been lost. It is important to know at least some mathematical details of the analysis to apply it properly in clinical practice. For all details about this case, see Supplementary Figure S1.

Figure 4

(A) A qEEG panel from a normal subject; from top to bottom: ADR (alpha-to-delta ratio) from the right and left hemispheres; RAV (relative alpha variability) from the major vascular territories: anterior cerebral artery (ACA), middle cerebral artery (MCA), posterior cerebral artery (PCA). (B) Awake with eyes open, (C) sleeping, and (D) awake with eyes closed. Note the variability on the ADR and on the RAV in a normal subject. (E) Fast Fourier transform from a period of EEG (rectangle on (D)); the power of each frequency band is shown in a different color. To calculate ADR, the power of alpha frequencies (8–13 Hz) is divided by the power of delta frequencies (1–4 Hz). To calculate RAV, the alpha power (8–13 Hz) is divided by the total power (1–20 Hz). ADR: alpha-to-delta ratio. RAVL relative alpha variability. L: left, R: right.

Figure 5

Two hours of a qEEG panel showing multiple focal seizures with the appearance of “solid flames”, one of which is highlighted with a shadow rectangle. (A) Rhythmicity spectrogram: y-axis frequencies from 1 to 25 Hz, showing an increase of rhythmicity (darker blue) in high frequencies at the beginning of the seizure that evolves to more rhythmicity in the lower frequencies at the end of the seizure, clearly more pronounced in the right hemisphere. (B) The spectrogram color scale represents power. Warmer colors (white–red) represent higher power and cooler colors (blue) represent lower power for each frequency band from 0 to 32 Hz (y-axis). Each seizure is a red–yellow flame more prominent in the right hemisphere. (C) Relative asymmetry spectrogram: a color scale represents power asymmetry in percentages between pair of homologous channels in both hemispheres at each frequency (y-axis from 1 to 18 Hz). White represents no asymmetry and the degree of darkness reflects more asymmetry until 50%; red marks more power in the right hemisphere and blue in the left hemisphere. In this case, there is a clear asymmetric increase in power in the right hemisphere during the seizures and because there is the presence of slow waves between seizures in the right hemisphere also, and asymmetry with an increase in power in the delta band is seen between seizures. Both amplitude trends, (D) amplitude integrated EEG (aEEG) and (E) peak envelop, show a consistent increase in amplitude during seizures, more accentuated in the right hemisphere. In this specific patient, the automated seizure detector recognizes the seizures; 10 s of raw EEG at (F) slowing in the background frequency, with the presence of lateralized periodic discharges (LPDs) over the right hemisphere at 0.4 Hz, with a plus modifier (superimposed fast activity). This pattern belongs to the ictal–interictal continuum. A total of 10 s of raw EEG at (G) continuous fast frequencies with spiky morphology with more amplitude in the right side, and diffusion to the left hemisphere. A total of 10 s of raw EEG at (H) the seizure has evolved, and rhythmic delta activity is present with superimposed spikes and sharp waves until the seizure abruptly finishes (green arrow). From Veciana and colleagues, 2024 [30] with permission.

Figure 5

Two hours of a qEEG panel showing multiple focal seizures with the appearance of “solid flames”, one of which is highlighted with a shadow rectangle. (A) Rhythmicity spectrogram: y-axis frequencies from 1 to 25 Hz, showing an increase of rhythmicity (darker blue) in high frequencies at the beginning of the seizure that evolves to more rhythmicity in the lower frequencies at the end of the seizure, clearly more pronounced in the right hemisphere. (B) The spectrogram color scale represents power. Warmer colors (white–red) represent higher power and cooler colors (blue) represent lower power for each frequency band from 0 to 32 Hz (y-axis). Each seizure is a red–yellow flame more prominent in the right hemisphere. (C) Relative asymmetry spectrogram: a color scale represents power asymmetry in percentages between pair of homologous channels in both hemispheres at each frequency (y-axis from 1 to 18 Hz). White represents no asymmetry and the degree of darkness reflects more asymmetry until 50%; red marks more power in the right hemisphere and blue in the left hemisphere. In this case, there is a clear asymmetric increase in power in the right hemisphere during the seizures and because there is the presence of slow waves between seizures in the right hemisphere also, and asymmetry with an increase in power in the delta band is seen between seizures. Both amplitude trends, (D) amplitude integrated EEG (aEEG) and (E) peak envelop, show a consistent increase in amplitude during seizures, more accentuated in the right hemisphere. In this specific patient, the automated seizure detector recognizes the seizures; 10 s of raw EEG at (F) slowing in the background frequency, with the presence of lateralized periodic discharges (LPDs) over the right hemisphere at 0.4 Hz, with a plus modifier (superimposed fast activity). This pattern belongs to the ictal–interictal continuum. A total of 10 s of raw EEG at (G) continuous fast frequencies with spiky morphology with more amplitude in the right side, and diffusion to the left hemisphere. A total of 10 s of raw EEG at (H) the seizure has evolved, and rhythmic delta activity is present with superimposed spikes and sharp waves until the seizure abruptly finishes (green arrow). From Veciana and colleagues, 2024 [30] with permission.

Figure 6

(A) One hour of a qEEG panel in a 67-year-old man with a past medical history of acute myeloid leukemia and renal failure. He was admitted with a fever (treated with cephepime) followed by confusion. The first part of the recording shows the presence of a broad band monotonous on the spectrogram and an increase in amplitude without asymmetry that matches with the raw EEG at (B) generalized periodic discharges at 1.7 Hz; this pattern belongs to the ictal–interictal continuum. The patient was confused and not following commands. Anti-seizure medication was administrated (intravenous clonazepam 1 mg, shadow rectangle). The raw EEG at (C) belongs to a period of patient clinical examination with muscle and movement artefacts on the EEG recording. Intermixed with the artefact, an improvement in the background EEG activity could be noticed. At that point, the patient followed commands. The raw EEG at (D) shows background slowing with no epileptiform discharges. The trend now shows a narrowband monotonous on the spectrogram, less amplitude on the aEEG, and peak envelope trends. The patient fulfils the criteria of nonconvulsive (electroclinical) status epilepticus with a clinical and EEG improvement of an ictal–interictal continuum pattern after anti-seizure medication. The rise in power and amplitude observed in the qEEG panel matches with artefacts as it is shown in (C).

Figure 6

(A) One hour of a qEEG panel in a 67-year-old man with a past medical history of acute myeloid leukemia and renal failure. He was admitted with a fever (treated with cephepime) followed by confusion. The first part of the recording shows the presence of a broad band monotonous on the spectrogram and an increase in amplitude without asymmetry that matches with the raw EEG at (B) generalized periodic discharges at 1.7 Hz; this pattern belongs to the ictal–interictal continuum. The patient was confused and not following commands. Anti-seizure medication was administrated (intravenous clonazepam 1 mg, shadow rectangle). The raw EEG at (C) belongs to a period of patient clinical examination with muscle and movement artefacts on the EEG recording. Intermixed with the artefact, an improvement in the background EEG activity could be noticed. At that point, the patient followed commands. The raw EEG at (D) shows background slowing with no epileptiform discharges. The trend now shows a narrowband monotonous on the spectrogram, less amplitude on the aEEG, and peak envelope trends. The patient fulfils the criteria of nonconvulsive (electroclinical) status epilepticus with a clinical and EEG improvement of an ictal–interictal continuum pattern after anti-seizure medication. The rise in power and amplitude observed in the qEEG panel matches with artefacts as it is shown in (C).

Figure 7

A computed tomography (CT) scan from an 81-year-old man with a cerebellar hematoma (A). (B) The quantitative EEG shows a progressive decrease in cerebral activity that it is revealed by a decrease in the power spectrum, changing from green (narrowband monotonous) to dark blue (lower power), as well as a decrease in the amplitude, as can be observed on the aEEG and envelope amplitude, along with an increase in the percentage of suppression. (C) The raw EEG at (C) shows a normal amplitude with the anterior–posterior gradient reversed. (D) The raw EEG at (D) shows a suppression of EEG activity corresponding to a low power spectrogram (dark blue), low amplitude on the aEEG and envelope train, and high suppression percentage. The patient was pronounced brain dead following confirmatory testing.

Figure 7

A computed tomography (CT) scan from an 81-year-old man with a cerebellar hematoma (A). (B) The quantitative EEG shows a progressive decrease in cerebral activity that it is revealed by a decrease in the power spectrum, changing from green (narrowband monotonous) to dark blue (lower power), as well as a decrease in the amplitude, as can be observed on the aEEG and envelope amplitude, along with an increase in the percentage of suppression. (C) The raw EEG at (C) shows a normal amplitude with the anterior–posterior gradient reversed. (D) The raw EEG at (D) shows a suppression of EEG activity corresponding to a low power spectrogram (dark blue), low amplitude on the aEEG and envelope train, and high suppression percentage. The patient was pronounced brain dead following confirmatory testing.

Figure 8

(A) Thirty minutes of a qEEG panel from a 71-year-old comatose man after resuscitation from a cardiac arrest. There are clear visible stripes on the qEEG trends, alternating between diffuse low power corresponding to suppression periods (raw EEG at (B)) and high power vertical stripes corresponding to bursts (raw EEG at (C)). On the raw EEG, the artefact reduction tool has been activated (ON) to remove the electromyography artefact (shadow grey in the raw EEG); then, EEG activity is clearly visible.

Figure 9

(a) qEEG in 2 h display from a 73-year-old man with past medial history of left ischemic stroke. In this panel is clearly visible, that seizure detector pointed 4 seizures (black arrows), however the seizure signature is clearly recognizable in many more occasions (orange arrows) that are not detected by the algorithm, but clearly recognizable using all the qEEG trends, cheeking the raw EEG all the arrows turn out to be seizures (raw EEG at (B) on Supplementary Figure S2 and matching clinically with eye version to the right and nystagmus. Rhythmicity spectrogram show a clear increase in high frequency bands at the beginning of the seizure that move to slow frequency bands, always more pronounced in left hemisphere. Spectrogram depicted a solid flame in the left hemisphere a little bit recognizable also in the right hemisphere. Asymmetry spectrogram is quite interesting; all the time shows a great power in low frequencies in the left hemisphere, on the raw EEG correlates with delta waves in that hemisphere, and during the seizures also an increase of left power in high frequency bands. aEEG and peak envelope depicted the typical arch shape in each seizure that tell us about and increase in amplitude during the seizures. In Supplementary Figure S2 you can find the raw EEG for the seizure pointed at (C). Note that after the intravenous antiseizure medication (ASM) administration the seizures separate and became better defined in evolution (a quite common situation in critical patients) therefore, easily recognizable by the automated seizure detector. First hour seizure burden was 16 min fulfilling the criteria of status epilepticus. (b) After the administration of lacosamide seizure burden decreases to 12 min per hour and seizures become cyclic. (c) Valproate was added and seizure burden further decrease to 5 min per hour, however a cyclic alternating pattern of encephalopathy also appear (blue line, raw EEG is shown in Supplementary Figure S2. Note the difference between cyclic alternating pattern more arch shape on the spectrogram (blue line) and seizures (red arrows) more triangular shape and smooth edges. Correlation with the raw EEG at (B) and (C) (CAPE), and at (D) (seizure) are shown in the Supplementary Figure S2.

Figure 9

(a) qEEG in 2 h display from a 73-year-old man with past medial history of left ischemic stroke. In this panel is clearly visible, that seizure detector pointed 4 seizures (black arrows), however the seizure signature is clearly recognizable in many more occasions (orange arrows) that are not detected by the algorithm, but clearly recognizable using all the qEEG trends, cheeking the raw EEG all the arrows turn out to be seizures (raw EEG at (B) on Supplementary Figure S2 and matching clinically with eye version to the right and nystagmus. Rhythmicity spectrogram show a clear increase in high frequency bands at the beginning of the seizure that move to slow frequency bands, always more pronounced in left hemisphere. Spectrogram depicted a solid flame in the left hemisphere a little bit recognizable also in the right hemisphere. Asymmetry spectrogram is quite interesting; all the time shows a great power in low frequencies in the left hemisphere, on the raw EEG correlates with delta waves in that hemisphere, and during the seizures also an increase of left power in high frequency bands. aEEG and peak envelope depicted the typical arch shape in each seizure that tell us about and increase in amplitude during the seizures. In Supplementary Figure S2 you can find the raw EEG for the seizure pointed at (C). Note that after the intravenous antiseizure medication (ASM) administration the seizures separate and became better defined in evolution (a quite common situation in critical patients) therefore, easily recognizable by the automated seizure detector. First hour seizure burden was 16 min fulfilling the criteria of status epilepticus. (b) After the administration of lacosamide seizure burden decreases to 12 min per hour and seizures become cyclic. (c) Valproate was added and seizure burden further decrease to 5 min per hour, however a cyclic alternating pattern of encephalopathy also appear (blue line, raw EEG is shown in Supplementary Figure S2. Note the difference between cyclic alternating pattern more arch shape on the spectrogram (blue line) and seizures (red arrows) more triangular shape and smooth edges. Correlation with the raw EEG at (B) and (C) (CAPE), and at (D) (seizure) are shown in the Supplementary Figure S2.

Figure 10

Cyclic alternating pattern of encephalopathy (CAPE): (A) cyclic alternating pattern is clearly visible on this EEG from an 82-year-old woman. There are like a series of arches on the qEEG trends, the top corresponding to a raw EEG at (B) showing generalized periodic discharges (GPDs) with triphasic morphology with amplitude around 100 µV, alternating with periods of theta background with amplitude around 50 µV (C). Cheyne-stokes respiratory pattern is correlated with the EEG changes, hyperpnea is present during high amplitude GPD and apnoea/hypopnea during low amplitude theta periods. When the medical team enter the room, the spontaneous alternating pattern shifts to a more amplitude and GPDs pattern, and the stimulation ((D,E) see Supplementary Figure S3D,E) induced the GPD pattern. GPDs pattern fulfils the criteria of IIC and in this specific patient, as it happens in many encephalopathic patients, correspond to the most stimulated state, despite the theta pattern looks like, apparently, more normal.

Figure 11

Normal (A1) quantitative trends and the (A2) raw EEG corresponding to the vertical arrow from a patient on day 2 after SAH. The same patient on day 6 after aSAH (B1) shows a decrease in the alpha-to-delta ratio in all territories, more pronounced in the left temporal regions. (B2) shows the raw EEG matching with the vertical arrow that corresponds to the moment after the stimulation showing slow waves more pronounced over the left hemisphere.