Natus & Persyst Instructions

Follow the setup steps, then use the Persyst tab to practice reviewing trends against raw EEG. Menus and saved presets vary by workstation version. Compare the display controls in EEG Fundamentals.

Natus Instructions

Launch & Log In

Power on the workstation, launch the Natus NeuroWorks app, and log in with your secure credentials.

Open Patient Study

Navigate to the Patient Studies list or use the search function to locate a record. Then select the desired EEG study. Confirm two patient identifiers, study date, and recording start time before interpreting or reporting.

Configure Display

Select the bipolar montage Bipolar LT over RT and begin with display sensitivity 7 μV/mm, time base 30 mm/sec, HFF 70 Hz, and LFF 1 Hz. The toolbar below also shows the 60 Hz notch filter, which removes mains (line) interference. These are starting display settings, not fixed interpretation settings; adjust the montage, sensitivity, time base, and filters to answer the clinical question while avoiding waveform distortion. See the same knobs on a 10-second page in EEG Fundamentals.

Calibrate the Screen

Go to Edit → Calibrate… On the Screen tab, set Monitor to Custom and enter the screen dimensions in millimeters using the actual measured display area, not a copied example value. Set Type to Auto (Primary), then click Apply. Calibration makes the on-screen millimeter scale physically accurate, so sensitivity (μV/mm) and time base (mm/sec) mean what they say.

Edit menu with Calibrate option highlighted

Calibrate dialog, Screen tab, Monitor set to Custom

Custom monitor dimensions with Type set to Auto (Primary)

Analyze the EEG

Read the raw EEG systematically: background, symmetry and reactivity, sleep features, artifacts, interictal abnormalities, and events. Change montage to check a suspected field; use video and ECG to test artifact explanations. Measure frequency and duration before assigning a pattern name.

Annotate Events

Place event markers by clicking the designated buttons at the top of the screen.

Create EEG Report

After review, open Epic's reporting interface, generate a new report, and pend the note for review.

Template Types (Epic SmartPhrases)

These SmartPhrases are institution-specific — they exist in this institution's Epic build and will not be available at other sites.

SmartPhrase	Use
`2026VEEG`	Standard VEEG (w/ prior routine in 6mo)
`2026VEEGNEONATAL`	Neonate VEEG
`2026EMUVEEG`	ALL EMU admissions (Kravis, MSH, MSW)
`2026BASELINE VEEG` Name to be confirmed: SmartPhrase names have no spaces.	New hookup without prior routine.
`2026SAHBASELINE…` Name to be confirmed.	Subarachnoid hemorrhage (alpha/delta ratio)
`2026INTRACRANIAL…` Name to be confirmed.	Intracranial studies (know the electrode locations)
`2026AMBULATORY`	Prolonged outpatient studies

Swipe sideways to see the whole table.

Persyst: Reading Quantitative EEG

How to use trends

Quantitative EEG (qEEG) turns hours of recording into a few strips of color: power at each frequency, how rhythmic the signal is, its amplitude, and how the two hemispheres compare. One screen can show 1–24 hours where the raw EEG shows ten seconds. Trends are for finding where to look. They do not replace looking.

Scan the whole panel at a long time base (2–6 hours) for anything that stands out from the background.
Flag each candidate: a flame, a band of rhythmicity, a jump in amplitude, a change in asymmetry.
Confirm on the raw EEG, two minutes either side of every flag. Only the raw EEG says whether it is a seizure, a periodic pattern, a state change or artifact.
Learn this patient's signature. Once one seizure is confirmed, its trend shape finds the rest, and lets you count the burden and follow the response to treatment.

What the evidence says

Spectrogram-guided review cut the time to review 24 hours of EEG from 38 to 8 minutes, with 87% sensitivity for seizures (Moura 2014).
Experts reading a qEEG panel together with the raw EEG found 63–68% of ICU seizures. An automated seizure detector found about 26%. Slow, low-amplitude seizures and seizures within periodic patterns were the ones missed (Haider 2016).
With a panel of rhythmicity, spectrogram, asymmetry and aEEG trends, EEG technologists and ICU nurses detected seizures about as well as neurophysiologists. No one should use the trends alone (Swisher 2015).

The trends, one by one

The small panels below are generated teaching sketches in the colors of the Persyst 15 trend screen, not screenshots. Check the color scales and settings on your own system.

Power spectrogram (CDSA, FFT)

Shows time across, frequency up (usually 0–20 Hz), power as color: warm is more power, cool is less. One strip per hemisphere.

Look for a sudden rise in power standing out from the background. Seizures often make a flame shape; sleep shows a spindle band near 12–14 Hz.

Watch out: color is relative to the scale chosen, and artifact adds power just as seizures do.

Rhythmicity spectrogram

Shows only the power that belongs to rhythmic, repeating activity (a Persyst trend). Pale is non-rhythmic; blue is strongly rhythmic.

Look for a band of rhythmicity that appears and drifts in frequency, as most seizures do.

Watch out: periodic discharges, rhythmic artifact (chewing, patting, ventilator) and sleep spindles are rhythmic too.

aEEG and suppression percentage

aEEG is amplitude after filtering and smoothing, plotted as upper and lower margins on a semi-logarithmic scale (linear to 10 µV, logarithmic above).

Look for the lower margin rising during a seizure, and dropping toward zero in burst suppression.

Suppression % is the share of each epoch below a small voltage threshold: about 0% for a continuous background, 50–99% in burst suppression.

Asymmetry spectrogram

Shows the difference in power between homologous left and right electrodes at each frequency. By convention blue is more power on the left and red more on the right.

Look for a focal seizure lighting up one side, then the opposite color as postictal suppression follows.

Watch out: a skull defect, scalp edema or one bad electrode makes a persistent asymmetry.

Alpha/delta ratio and alpha variability

Shows alpha power divided by delta power per region, trended over days. Relative alpha variability grades how much alpha fluctuates.

Look for a sustained regional fall after subarachnoid hemorrhage. A post-stimulation ADR falling more than 10% for six consecutive readings, or more than 50% once, was associated with delayed cerebral ischemia (Claassen 2004).

Watch out: sedation, sleep and artifact move it too; compare with the patient's own baseline.

Seizure probability and detections

Shows how likely an algorithm judges a seizure to be, moment by moment (the black line), and where it called one (red blocks, which hide the line).

Look for clusters of detections that line up with flames and rhythmicity.

Watch out: detectors miss low-amplitude, slow and focal seizures and seizures within periodic patterns, and alarm on artifact. Treat each detection as a prompt to read the raw EEG.

vsBaseline spectrogram

Shows each moment's power compared with this patient's own baseline, frequency by frequency (a Persyst trend; the baseline is chosen automatically unless you set it). White is no change, magenta more power than baseline, teal less. The sketch is the right hemisphere over three days after subarachnoid hemorrhage: alpha fades (teal) and delta grows (magenta).

Look for slow changes over hours or days that an ordinary spectrogram hides in its background, and seizures standing out as blocks of magenta.

Watch out: the comparison is only as good as the baseline. If the baseline was taken during sedation, a seizure or artifact, everything after it looks changed.

Spectrogram patterns and their names

A proposed vocabulary for spectrogram patterns helps teams describe what they see. Each name points to what the raw EEG usually shows, never to a diagnosis on its own.

Solid (regular) flame. An abrupt rise in power with smooth edges that stands out from the background: the most recognizable seizure pattern.

Choppy (irregular) flames. Brief, ragged bursts of power, usually state changes or arousals rather than seizures.

Broadband monotonous. Sustained high power across more than 5 Hz of bandwidth: ongoing status epilepticus or continuous periodic discharges.

Narrowband monotonous. Power confined to the slow frequencies (below 5 Hz) with little change: typical of encephalopathy.

Stripes. Vertical bands where power alternates with suppression: burst suppression or burst attenuation.

Suppression (low power). Continuously low power across frequencies: a suppressed background, or a disconnected or shorted electrode.

Repeating flames. Seizures recurring at fairly regular intervals (sometimes called cyclic seizures): count them to measure the burden.

Artifact. Not a pattern name, but worth knowing: vertical bands that start and stop abruptly, mostly in the fast frequencies, on both sides at once, without evolution.

Trend lab

Six situations from ICU monitoring. Read the panel, click anything that stands out to check the raw EEG underneath, then choose your reading. The panel is laid out like the Persyst 15 trend screen with the vsBaseline Comprehensive template; switch to Comprehensive from the template menu. The data are generated teaching examples, not recordings.

The trend lab needs JavaScript.

Set up Persyst in NeuroWorks

Menus, templates and color scales vary between versions and sites. The pictures are drawn after Persyst 15 with generated teaching data, not screenshots.

Launch Persyst

Once your EEG study is open, go to Analysis → Other Tools and make sure Persyst (toolbar) is checked.

Persyst (toolbar)

Locate the Persyst toolbar as pictured below. If using two screens, grab the toolbar and drag it to the second screen.

Drawn after the Persyst 15 toolbar; the version on your station may differ.

Configure Panel Size

Check that a Comprehensive template is selected (here vsBaseline Comprehensive). Adjust display size by grabbing the bottom right corner to expand the panel.

Drawn after the Persyst 15 trend screen, with generated teaching data; not a screenshot.

Rhythmicity Spectrogram

The rhythmicity spectrogram highlights sustained rhythmic activity across frequencies and channels. It is a screening display: seizures, rhythmic/periodic patterns, artifact, and normal rhythms can all produce conspicuous trends, so every candidate event must be confirmed on raw EEG.

FFT Spectrogram

The FFT spectrogram displays frequency power over time. Warm colors indicate relatively greater power and cool colors lower power on the selected scale; color alone does not distinguish seizure from artifact, sleep, medication effect, or other cerebral activity. Confirm meaningful changes on raw EEG.

Asymmetry and aEEG

The Asymmetry panel in Persyst uses color contrasts (left blue and right red) to reveal lateral differences in EEG power between hemispheres sorted from delta to gamma frequency on the Y axis. The aEEG panel compresses long-term amplitude trends into a waveform, together providing a quick visual summary of lateral power differences and amplitude trends. Check the displayed color legend and scale for the selected template. An asymmetry does not identify its cause; confirm it on raw EEG and inspect electrode quality.

ADR

The alpha/delta ratio (ADR) trends relative alpha and delta power. A sustained regional decline can provide an early warning of delayed cerebral ischemia after subarachnoid hemorrhage, but an isolated value is neither “normal” nor diagnostic. Review baseline, raw EEG, asymmetry, sleep, sedation, artifact, systemic physiology, and the clinical examination.

Pitfalls

Trends lag and average. Each pixel summarizes several seconds, so brief events and low-amplitude seizures can vanish.
Seizures within periodic patterns are hard to see, because the background already has high power and rhythmicity. Read the raw EEG when the spectrogram is broadband monotonous.
Focal seizures may show only on a regional trend. Once you know where they start, build a trend from those electrodes.
Artifact looks like power. Chewing, patting, suction, ventilators and ECG all add power; abrupt onset and offset without evolution are clues.
One bad electrode makes an asymmetry. Check impedances and the raw EEG before calling a hemisphere abnormal.
Color scales are relative. Keep the same template and scale when comparing across hours or days.
Trends are not reports. Describe what the raw EEG shows in ACNS terms, and use trends to count burden and show change over time.

Sources: Veciana de las Heras et al., Brain Sciences 2024 (qEEG in the ICU with Persyst; spectrogram terminology) · Moura et al., Neurology 2014 · Haider et al., Neurology 2016 · Swisher et al., J Clin Neurophysiol 2015 · Claassen et al., Clin Neurophysiol 2004 · Rosenthal et al., Ann Neurol 2018 · Bitar, Khan and Rosenthal, Crit Care 2024