SKIL TOPOMETRIC
SOFTWARE MANUAL

1. Create Report Click Blank Page icon SKIL starts with a blank canvas. A common mistake is attempting to open an EEG file now, before a new report is created. This will fail. Most functions are unavailable (icons ghosted), including opening or loading an EEG file. Only two functions can be performed now: Create Report [blank page] or Open Report [open folder].

Dialog window opens. Creating or opening a report opens a standard dialog window. Navigate to the folder where you plan to store (or open) the new report.

Enter name, create. Enter a file name and press Create to make new report.

 

2. Report Explorer Reports are organized around the Report Explorer, a central index, which guides you in completing all elements of a QEEG evaluation. This includes 1) demographic information, 2) acquisition methodology, 3) summary of findings, and 4) conclusions and recommendations. Conclusions and recommendations can be supported by brain maps, topometrics, statistical tables, spectral plots, comodulation maps, and samples of typical and atypical EEG data, all of which can be incorporated into the final report. The Report Explorer lists all possible elements of a report: text, EEG charts, and figures. Click on a element name in the Tree View (left) to open its Contents window (right). Click on a folder to reveal or hide elements.

3. Open EEG File Click the Load EEG icon. This opens a standard dialog window: navigate your hard drive to where your EEG files are stored. (You may have trouble accessing files on Zip disks or CD-ROMS. If so, simply copy EEG files to a hard drive folder.)

4. Process Data File When an EEG data file is loaded, its Chart appears, displaying the initial few seconds of data plus a set of tools. Visual inspection, artifact detection and deletion, and data selection for analysis are performed here. Chart tools can alter how an EEG record is displayed. Common monopolar and bipolar montages are available (the default "Ref" = 19 sites, referential). Display gain (arbitrary height of signals) and display length can be optimized by the user. Scale is tied to display gain and indicates the microvolt height of the cursor. Place the cursor adjacent to any waveform whose amplitude you want to estimate. 4a. Data Selection: To select data segments for analysis, click anywhere within the chart, hold down the mouse button, and drag right. A red line is drawn in the status area above the chart; the segment of data below it has been "selected" and will be spectrally analyzed when the file is processed. To mark artifacts in the record, repeat the click, hold, and drag while also holding down either Shift key. A purple line is drawn in the status area. Purple lines indicate artifact and are excluded from analysis; red lines indicates selections to be analyzed. Purple overrides red: segments marked with both colored lines are excluded from analysis. To erase or amend a selection, simply click and drag left (hold down the Shift key to erase artifact marks). To erase all selections from a record, select Clear Sel in Edit menu. You can select any number of data segments. To select entire record, click Select All .

5. Artifact Detection and Removal For most reports, de-artifacting will proceed in two stages: 1. automatic detection followed by 2. visual inspection. During visual inspection you can amend artifact marks (purple lines) laid down by the automatic process and identify and mark any artifacts that were missed. Our detection algorithm evaluates waveforms from a single channel (e.g., Fp1 or X1, the first auxiliary channel, often an EOG lead). It is pre-set to evaluate channel Fp1 with a 3000 amplitude threshold. Click Mark Artifact on the Mark Artifact Menu to process record for artifacts. You can make the detection process more or less sensitive by decreasing or increasing the amplitude value, respectively. Experiment with the amplitude value, clicking Mark Artifact after each change to witness the outcome. Click Done when you're satisfied with automatic identification. You can always amend any mark manually, and add or delete marks entirely by hand. Be sure to inspect the entire record for any artifacts missed by the automatic process. Also, to minimize transition effects, the first 30 s of all records are ignored (marked as artifact). You can set the initial ignore to any length, including zero.

6. Create Figure Any images of raw and spectral data can be captured and included in the final report as figures. Click the Append (camera lens) icon to capture the image of whatever window is currently forward -- EEG Chart, Grid View, Brain Map, etc. When the Append icon is clicked, the Caption Menu opens. Provide a unique figure name and enter an image description (condition and other relevant information). You can create figures and later decide to exclude them from the report.

7. Process EEG Click Process EEG to start analysis of your data selections. In the Spectral Analysis menu indicate what condition these data selections represent by clicking the appropriate radio button. If the EEG files' timestamp is inaccurate (e.g., past PC clock error, daylight savings) the Time of Day can be corrected here. Click OK to submit this data to spectral analysis.

8. Available Views and Tools Conditions Frequency Bands View Modes Data (client means or % change) or Stat (z-scores relative to database). Time of Day Correction: If checked, database comparisons are corrected for time of day. Not available for tasks.

9. Grid View Grid view presents spectral means in tabular format. In Stat view, values greater than +2 SD or less than -2 SD are bold.

10. Spectral Plot Spectral plot displays spectral magnitudes in standard 1-Hz arrays. Display gain and spectral range can be adjusted. In Stat view, only values above a specified standard deviation value appear (in red). Spectral plot - Client means

Spectral plot - Deviations from Database

11. Brain Maps Brain maps, also called topographs, display spectral magnitudes by means of color variation. Head size and scale are adjustable. To capture an image of a brain map, you may want to optimize it prior to clicking Append. Do this by dragging the right window corner around with the mouse. For clinical bands you may want to increase head size and manually resize the window to create two rows of three heads each. If the heads are too large and rows too many (e.g., 1 Hz figures), the figure may not fit the page. The scale can vary greatly from map to map so be sure to optimize it accordingly for each map.

12. Topometrics A topometric displays spectral magnitudes, relative differences, or statistical results for every site for a single frequency band. Right-click anywhere to access customization controls. Single State Topometric State Comparison Topometric

Multiple-State Topometric

13. Comodulation Comodulation maps depict functional homogenity and topographic reciprocities within a single frequency. Each head displays comodulation (rho) values between one recording site and all areas of the head. Display gain and head size are adjustable. In Data View, each recording site correlates perfectly with itself (1.0), of course. The default band is 8-12 Hz. Use Edit Band in the Edit Menu to change the frequency band.

14. Replications When a 2nd (3rd, 4th, etc) EEG file is processed for the same condition, the Spectral Replication menu appears, asking whether the current file is a new replication or a revision of a previous replication. You must indicate which. Click Spectral Results folder of the Report Explorer to manage replications. Click the appropriate enable button to include a replication in the current views. When one enable box is checked, only this recording is viewed. When two or more enables are checked, replications are averaged together in each view. The exception is for Topometrics, where a "replicates" box can be checked. When is checked, replications are displayed separately (up to 6 reps).

15. Topometric Replications As stated above, replications are managed at the Spectral Results folder of the Report Index Topometrics from the first replication is presented in the top figure, and from the second replication in the bottom figure.

 

Replications are averaged in the top figure and presented individually in the bottom figure.

16. Customize Frequency Band Specify any frequency band by clicking Select in the Band List, then access Edit Band in the Edit Menu. Enter low and high frequency ranges, provide a label, and press OK.

17. Report Finalization and Printing Review the various figures you have collected from the client's data. From this information complete Findings and Conclusions. Figure Controls allow you to exclude figures from a report or to reorder them. Right-click on the figure name or movement box (box with circle). Select MoveUp or MoveDown to move figure one position. To move a figure further, repeat the process. To exclude a figure, uncheck the box. To delete a figure entirely from the report file, click the name and hit Del on your keyboard.

 

18. Save Report Open the Report Explorer and select Client Info. This will make the Save icon (floppy disk) unghost. Clicking the icon now saves the report. Reports saved in this fashion should be restorable after program crashes. You can save a report at any time during its creation.

19. Print Preview IMPORTANT: Do not use the Print command in the File Menu (it will be removed in future versions). In order to print a report, first select Print Preview. Print preview provides a review of the entire report. Scan through the report for errors, figures accidentally included or excluded, etc. Cancel the preview and return to Report Explorer to make corrections or additions. When you're satisfied with the report, press Print!

20. Printed Report A final report can be any length. The first two pages contain client information and recording session methodology; additional pages contain captioned images and your conclusions. The last page provides information about each recording (usually for your records only).

 

Appendix Adult Database Description The adult database was derived from laboratory studies of 135 human subjects ranging in age from 18 to 55 years. Approximately 80% were males and 20% females. This population consisted of students and laboratory personnel (50%), recruited volunteers from the community (25%), and U.S. Air Force personnel, including pilots, ground crew, and administrative staff (25%). All subjects completed a handedness inventory (Oldfield, 1971) and a comprehensive questionnaire designed to provide a screen for medical history, drug use, and recent life events. All subjects reviewed and signed an approved institutional information and consent form. Air Force personnel were also intensively pre-screened as a condition of their service, and were subject to regular medical examinations and an unusually high level of drug-use scrutiny. The data recorded from these subjects included 2- to 3-minute eyes closed, eyes open, and engagement (task) conditions. Engagement tasks were 1) audio-visual information processing and 2) a visuo-spatial tracking. Each subject provided two to four replications per condition. Since it was often the case that these replications were spaced by several hours, or in some cases days, it was possible to obtain a combination of cross-sectional and longitudinal data to fill out the time intervals used to evaluate diurnal influences on EEG characteristics. Twenty-one time-of-day categories were established using 274 eyes closed and 274 eyes open conditions across the day. Each categories corresponds to a one-hour interval, spaced every half hour from 9 AM to 7 PM (e.g., 8:30 AM-9:30 AM, 9:00 AM-10:00 AM). The number of subjects in each interval ranged from 15-19 in the morning and evening hours to 29-38 from noon to 5 PM. No subject was sampled more than once in a time interval. Comparisons to either active task did not require time-of-day correction. Time-of-day correction for inactive (eyes closed, eyes open) conditions is highly recommended. Eyes closed and eyes open data show prominent circasemidian (about a half day) and ultradian (about 90-120 min) cycles (Kaiser & Sterman, 1994). Numerous papers in the EEG literature confirm these observations (Cacot et al., 1995; Frank et al., 1966; Gundel and Hilbig, 1983; Gundel and Witthoft, 1983; Machleidt, 1980). There is also an extensive literature in the sleep field showing a clear circasemidian cycle of sleep propensity, and reciprocally, attention, as measured by task performance (for excellent reviews see Broughton, 1995, and Broughton and Mullington, 1992). This literature discloses a transitory afternoon time zone characterized by increased sleep propensity and degraded performance in attentional tasks. This time period corresponds exactly to the mid-day increase in many frequency components documented by our database. Studies have shown that this "post-lunch dip", as it is sometimes referred to, is clearly independent of food intake! Collectively, these literatures and our own empirical findings establish the critical importance of a time-of-day correction in database construction for QEEG analysis.