This method can help answer key questions in the cognitive neuroscience field. The main advantage of this technique is that it can analyze neural activity and connectivity on high-spatial temporal evolutions. To begin analysis of intracranial EEG data, set up SPM12 and select the MEG EEG analytical menu.
First, perform time-frequency analysis for the preprocessed intracranial EEG data of each trial using continuous wavelet decomposition with Morlet wavelets based on predefined parameters. To reveal the temporal evolution of spectral components, conduct wavelet decomposition using seven-cycle Morlet wavelets for the entire epoch of 1, 000 to 2, 000 milliseconds using the frequency range of four to 300 hertz. Next, determine the mother wavelet and the number of cycles and note that the number of cycles controls the time-frequency resolutions and should be greater than five to ensure estimation stability.
Determine time and frequency ranges. Then crop the resultant time frequency maps automatically to remove edge effects. Here, the maps are cropped at 200 to 500 milliseconds.
Perform data transformation if desired and baseline correction by selecting time frequency rescale for the time frequency maps to better visualize event-related power changes and improve normality. Next, convert the time frequency maps into two-dimensional images. Also smooth using a Gaussian kernel with a predefined full-width half-maximum value to compensate for inter-subject variability and to conform to assumptions of random field theory.
Now select specify second level in the SPM menu and enter the 2D images that will be analyzed. Then run the general linear model by selecting model estimation. Finally, select results to perform statistical inferences on the time-frequency data based on random field theory.
Detect significantly activated time-frequency clusters with predefined thresholds such as those seen here. Start dynamic causal modeling analysis by selecting DCM in the SPM menu. Then choose the IND option and select new data to import the preprocessed intracranial EEG data.
Next, use the MEG EEG menu to specify the time window of interest, frequency window of interest, number of wavelet cycles that will be used, conditions of interest, and the contrasts for the conditions. Set the time window to one to 500 milliseconds. Use five-cycle Morlet wavelets of four to 100 hertz in one hertz steps.
Use the default setting for the wavelet cycle. Determine the time-frequency ranges based on research interest. Note that a time window with an additional 512 milliseconds can be automatically used during computation to remove edge effects.
Based on the DCM framework, define the driving inputs which represent sensory inputs on neural states and the intrinsic connections which embody baseline connectivity among neural states. Also, define the modulatory effects on the intrinsic connections via experimental manipulations for null and hypothesized models. Define the type of modulation as linear or nonlinear.
Now specify the intrinsic linear and nonlinear connections, driving inputs and modulation inputs. Modify the default settings of related parameters if necessary such as prior stimulus onset time and duration. Then choose invert DCM to estimate the models.
After that, select save results as image to save frequency-frequency modulatory coupling parameter images. Next, conduct a random effects Bayesian Model Selection analysis by selecting BMS to identify the optimal network model. Use the model expected probabilities and accedence probabilities as evaluation measures.
Then make inferences regarding the cross-frequency patterns of the modulatory connections using the winning model parameters. Now smooth the modulatory coupling parameter images by selecting convert to images. Then use specify second level to perform general linear model analysis.
Finally, select results to calculate the 2D SPMT values. Here, the full-width half-maximum was set at eight hertz and significant values were exploratorily identified using an uncorrected height threshold of P less than 0.05. Time-frequency analyses were conducted to investigate the temporal and frequency profiles of Inferior Occipital Gyrus or IOG activity during the processing of phases.
Here we see time-frequency maps of the right IOG activity for the upright phase and upright mosaic conditions. The SPMT data for upright phase versus upright mosaic are also shown. Functional network models are shown here.
Eight possible combinations of modulatory input of upright phase versus upright mosaic onto connections between the IOG and amygdala and self-connection onto each region were investigated. Frequency-frequency modulatory coupling parameters and SPMT values for upright phase versus upright mosaic for the IOG versus amygdala and amygdala versus IOG modulation are shown here. The red-yellow areas indicate significant excitatory connectivity while the blue-cyan areas indicate inhibitory connectivity.
After watching this video, you should have a good understanding on how to analyze intracranial EEG data for detecting neural activity and connectivity using the SPM software.