ABSTRACT
An understanding of spike fields is critical for accurate interpretation of the EEG. We developed a computer simulation tool that takes a user-defined scalp potential distribution as input and produces the associated EEG spike-wave complex in longitudinal bipolar, transverse bipolar, and referential montages simultaneously. Users choose single or multiple foci of maximum potential on a 2-dimensional electrode map to create EEG spikes with fields of variable complexity on an organized user-adjustable background. Distances between electrodes were determined by their coordinates in 3-dimensional space, and used to calculate normalized voltages that spread according to an exponential decay function. The length-constant used for the decay function can be adjusted by the user to manipulate the scalp potential spread and size of the EEG spike field. Using this simplified model, the simulation successfully translates a scalp potential input into the expected EEG. This simulation would be useful both as a teaching tool and for interpretation of EEG spikes with complex fields.
BACKGROUND
Scalp EEG is a tool for measuring electrical activity generated within the brain. EEG serves an important role in localization of ictal and interictal sources. Localization is guided by analysis of phase reversals and fields of spikes and sharp waves. Knowledge of typical and atypical patterns of spike fields is necessary for accurate EEG interpretation. This interactive model is primarily designed as a teaching tool for aiding EEG analysis. The model does not explain source location within the brain, but instead generates predicted EEG patterns given a scalp potential distribution as input.
METHODS
Electrode map. Electrode locations were based on MRI measurements [1]. Mouse clicks were mapped to a 40x28 array, overlaid on the scalp image. Heatmap colors were assigned based on normalized voltages between 0.5 and 1.
Voltage distribution. The location of a click was normalized to 1. Array values were determined according to the decay function, Vd = Vmax ⋅ e-d/λ, where distance d was calculated based on published x,y,z coordinates for the electrodes [1]. After electrode coordinates were fit to the array, linear interpolation of x,y values and weighted averages of z values were used for inter-electrode locations, thus approximating the shape of the scalp in three dimensions.
EEG plot. Values of relevant electrode locations on the array were subtracted and plotted as a spike-wave for each trace on all montages simultaneously. Slow-wave amplitudes were fixed proportional to spike amplitude. Background rhythms were generated using sine functions based on user frequency settings with randomized offsets.
INSTRUCTIONS
- Click on single or multiple locations on the scalp image to define areas of maximum voltage.
- To adjust the spread of the voltage distribution, change the value of the "Decay factor", and click "RESET/UPDATE".
- To change the number of spikes plotted, change the value of "Spike #", and click "RESET/UPDATE".
- To increase the sensitivity of the EEG, decrease the value of "Sensitivity" (μV/mm), and click "RESET/UPDATE". To decrease sensitivity, increase the value.
- To adjust the background of the EEG, use the dropdown menus to specify values for generalized delta, generalized theta, posterior alpha, and frontocentral beta.
- To create a custom montage, use the dropdown menus next to the third EEG plot to specify which two electrodes to subtract for each trace.
KNOWN ISSUES
- ***If the window has been scrolled down or to the right, the spikes will be plotted erroneously. If the heatmap does not align with the click location, make sure the window is scrolled completely to the top and to the left.***
- Occasionally, the heatmap distribution appears non-uniform. This is a result of (1) using a 2-dimensional image to represent a non-spherical 3-dimensional structure, and (2) imperfect estimation of x,y,z coordinates between electrode locations.
- Internet Explorer is not supported.
- This simulation has been tested successfully on later generations of iPads, but does not work correctly on older iPads.