CN116269441A - Epilepsy area positioning method based on high-frequency oscillation and connectivity - Google Patents

Abstract

The invention discloses a method for positioning an epilepsy induction zone based on high-frequency oscillation and connectivity, which comprises the following steps: 1. collecting SEEG data and selecting a representative channel; 2. filtering to obtain 54-200Hz signals and selecting baseline data and target data; 3. acquiring normalized high-frequency energy; 4. calculating a time coefficient and an energy coefficient and obtaining a high-frequency epilepsy index; 5. filtering to obtain a 12-45Hz signal; 6. calculating nonlinear regression analysis; 7. calculating the total intensity; 8. the connectivity high frequency epilepsy index was defined and calculated and performance was assessed. The invention can accurately position the epileptogenic region of patients with different epileptic seizure patterns and has potential application value in epileptic treatment.

Description

An Epileptogenic Zone Localization Method Based on High Frequency Oscillations and Connectivity

technical field

The invention relates to the field of multi-channel neural signal analysis of epileptic patients, in particular to high-frequency oscillations and a correlation index based on brain network connectivity, which are used to locate the epileptogenic zone of patients with different epileptic seizure patterns.

Background technique

Epilepsy is a serious, chronic neurological disorder that disrupts the normal activity of neurons in the brain. During the seizure, the patient may be injured or even life-threatening in an emergency, which brings great psychological pressure and difficulties in work and life to the patient. For people with drug-resistant epilepsy, epilepsy surgery can effectively treat seizures. The epileptogenic zone is defined as the brain region where the seizure occurs. Typically, the assessment of the epileptogenic zone involves multichannel intracranial electroencephalographic (iEEG) recordings, particularly stereoscopic electroencephalography (SEEG). How to identify the epileptogenic zone through these records is very important before surgery, because the purpose of surgery is to remove the epileptogenic zone. Indeed, precise localization of the epileptogenic zone has been a major challenge in epilepsy surgery.

In recent years, researchers have developed methods such as epileptogenicity index (EI), epileptogenicity map, high-frequency epileptogenicity index (HFEI), epileptogenicity grade (ER), and epileptogenic zone fingerprint to locate epileptogenic zones. These methods, based on the detection of the time-frequency or spatial properties of high-frequency oscillations to locate the epileptogenic zone, have received a lot of attention and achieved remarkable results in many studies, but they cannot be well localized in some cases epileptogenic zone. In epilepsy patients, approximately 75% of seizure patterns involve high-frequency oscillations (HFOs), while the remainder are slow-seizure patterns. The efficiency of EI, HFEI, and other methods based on HFO detection will become lower for the slow-onset mode. On the other hand, when similar seizure activity is presented in the recording channels, these methods will not be able to distinguish their temporal order or energy intensity differences, resulting in poor estimation results.

In general, epilepsy can be considered a network disorder, and HFO cannot capture the network properties of the brain by processing each channel individually. That is, relying solely on the electrophysiological characteristics of HFOs is not sufficient when HFOs are not observed in the ictal zone. On the other hand, previous research has shown that removing hyperexcitable sites is not the best way to reduce seizure rates. Instead, it is often more efficient to remove normal sites located at key points of the network, i.e. "drivers", according to the concept of network structure and connectivity. This is because the connectivity between SEEG signals increases before seizures, gradually decreases during early seizures, and gradually increases during late seizures.

Such as: Wang Haixiang. The application of stereotaxic EEG epilepsy index analysis in the location of epilepsy zone and the evaluation of epilepsy network [J]. Chinese Journal of Neurology. 2017; 6:362-367. This paper proposes a new Based on the SEEG quantitative method of stereotaxic electroencephalography (SEEG) analysis of high-frequency activity (60-90Hz) during seizures, the high-frequency epileptogenic index (HFEI) is calculated to locate the epileptogenic zone and evaluate the epileptogenic network in epileptic patients. However, this paper only used high-frequency oscillations of 60-90Hz, ignoring the slow onset mode of epilepsy, which would bring poor predictive effect when locating the epileptogenic zone of epilepsy patients with slower onset mode.

Contents of the invention

In order to solve the above technical problems, the present invention provides an improved epileptogenic zone location method - connectivity high-frequency epileptogenicity index (cHFEI), which combines the connectivity of HFO and brain network, and utilizes the connectivity of HFO and brain network Connectivity to locate the epileptogenic zone more accurately, with accurate positioning and high stability.

In order to solve the problems of the technologies described above, the technical solution adopted in the present invention is:

A method for locating the epileptogenic zone based on high-frequency oscillation and connectivity, characterized in that it comprises the following steps:

Step S1: collecting SEEG data and selecting representative channels;

Step S2: Filter and acquire 54-200Hz signal and select baseline data and target data;

Step S3: Obtain normalized high frequency energy NHFE;

Step S4: Calculate the time coefficient TC and the energy coefficient EC according to NHFE, and obtain the high-frequency epileptogenicity index HFEI;

Step S5: filtering to obtain a 12-45Hz signal;

Step S6: calculating nonlinear regression analysis;

Step S7: Calculate the total intensity TOT;

Step S8: Define and calculate the connectivity high-frequency epileptogenicity index cHFEI, and locate the epileptogenic zone according to the cHFEI.

The further improvement of the technical solution of the present invention lies in: in the step S1, according to the method of stereotaxy, collect the three-dimensional electroencephalogram SEEG data of twenty patients with different epileptic seizure patterns, and select 10- 30 representative channels covering the main area of the seizure.

The further improvement of the technical solution of the present invention lies in: in the step S2, use the second-order IIR notch digital filter and the fifth-order IIR Butterworth digital filter to filter the original signal to the high-frequency band of 54-200Hz; manually select the baseline Data and target data, the principle of baseline data selection is SEEG data with no obvious abnormal signal before the seizure, and the target data needs to cover the entire seizure process.

The further improvement of the technical solution of the present invention lies in: in the step S3, the bandpass signal is converted into a high-frequency energy spectrum by means of amplitude square and window smoothing, and the average value of the high-frequency energy of the baseline data is calculated as the high-frequency energy Baseline value; divide the high-frequency energy of all channels by the high-frequency energy baseline value of this channel to obtain the normalized high-frequency energy NHFE.

The further improvement of the technical solution of the present invention is that: the specific operation of the step S4 is:

Calculate the time-to-onset threshold for each channel i, defined as the maximum value of the baseline NHFE plus ten times the standard deviation of the baseline NHFE:

threshold _{onset time} = max(NHFE _BL )+10σ(NHFE _BL ),

When the normalized energy of the target data of each channel exceeds the threshold, this moment is judged as the time when the abnormal activity starts, and the channels are sorted according to the starting time of each channel, and the time coefficient TC is defined as the reciprocal of the order of each channel ;

Use NHFE to calculate the average energy within 250ms before and after the earliest start of each channel as the energy coefficient EC, and finally get the HFEI of each channel i:

The further improvement of the technical solution of the present invention lies in: in the step S5, band-pass filtering is performed on the original data, the pass band is β-γ frequency range, 12-45 Hz, and then down-sampled to 256 Hz.

The further improvement of the technical solution of the present invention is: in the step S6, calculate the nonlinear regression analysis based on the ^h2 nonlinear correlation index, perform segmented linear regression between each pair of signals, and test a signal relative to the other within the maximum lag All offsets of a signal; for two signals x and y, the piecewise linear approximation f of the transfer function between x and y is defined as:

y _n-τ = f(x _n )+e _n ,

^h2 measures the goodness-of-fit of the regression, explained by approximating f:

The further improvement of the technical solution of the present invention is: in described step S7, calculate all paired h ² values, use length is 3s, the sliding window that step size is 1s, maximum delay is 0.1s, produces a binary connection matrix; Then In each time window, for each channel, the orientation-independent total intensity TOT is summed.

The further improvement of the technical solution of the present invention is: in the step S8, for the SEEG channel of each patient, the average value and HFEI of the total intensity 8s before the onset are summarized, and normalized respectively, combined with TOT and HFEI, to obtain The composite index cHFEI of the Connective High Frequency Epileptogenicity Index:

The further improvement of the technical solution of the present invention lies in: the positioning of the epileptogenic zone is accurate and the stability is high.

Owing to having adopted above-mentioned technical scheme, the technical progress that the present invention obtains is:

The invention can locate the epileptogenic zone more accurately, and solves the problem that the accuracy of the positioning method based only on HFO is not high in the slow seizure mode.

The present invention solves the problem that the time sequence (or onset delay) or energy intensity of each channel cannot be detected when the positioning method based only on the time-frequency characteristics of HFO shows similar activities in the channels, resulting in low accuracy in locating the epileptogenic zone. high question.

Description of drawings

Fig. 1 is a flowchart of the present invention;

Figure 2 is a comparison chart of the area under the ROC curve (AUC) of different methods.

Detailed ways

Below in conjunction with embodiment the present invention is described in further detail:

The present invention is a method for locating an epileptogenic zone based on high-frequency oscillation and connectivity, comprising the following steps, as shown in Figure 1:

Step S1: Acquire stereoscopic electroencephalogram (SEEG) data of twenty patients with different seizure patterns according to the stereotactic method, and select 10-30 channels according to the seizure status of the patients, covering the main areas of seizures .

Step S2: Filter the original signal to a high frequency band of 54-200 Hz by using a second-order IIR notch digital filter and a fifth-order IIR Butterworth digital filter. Baseline data and target data were manually selected. The principle of baseline data selection was SEEG data without obvious abnormal signals before seizures, and the target data should cover the entire seizure process.

Step S3: Convert the bandpass signal into a high-frequency energy spectrum by means of amplitude square and window smoothing, and calculate the average value of the high-frequency energy of the baseline data as the baseline value of the high-frequency energy. Further, divide the high frequency energy of all channels by the high frequency energy baseline value of this channel to obtain normalized high frequency energy (NHFE).

Step S4: Calculate time coefficient (TC) and energy coefficient (EC) according to NHFE, and calculate high frequency epileptogenicity index (HFEI) of each channel.

Calculate the time-to-onset threshold for each channel i, defined as the maximum value of the baseline NHFE plus ten times the standard deviation of the baseline NHFE:

threshold _{onset time} = max(NHFE _BL )+10σ(NHFE _BL );

When the normalized energy of the target data of each channel exceeds the threshold, we judge this moment as the time when the abnormal activity starts, and sort the channels according to the starting time of each channel. The time coefficient TC is defined as the reciprocal of the sequence of channels .

Use NHFE to calculate the average energy within 250ms before and after the earliest start of each channel as the energy coefficient EC. Finally get the HFEI for each channel i:

Step S5: Perform band-pass filtering on the original data, the pass band is the β-γ frequency band (12-45 Hz), and then down-sample to 256 Hz.

Step S6: Calculating a nonlinear regression analysis based on the ^h2 nonlinear correlation index, performing a piecewise linear regression between each pair of signals, testing for all shifts of one signal relative to the other within a maximum lag. For two signals x and y, the piecewise linear approximation f of the transfer function between x and y is defined as:

y _n-τ = f(x _n )+e _n ,

^h2 measures the goodness of fit of the regression and is equivalent to ^r2 used in linear regression, which can be explained by approximating f:

Step S7: Calculate all pairs of ^h2 values, using a sliding window of length 3s, step size 1s, and a maximum delay of 0.1s, resulting in a binary connectivity matrix. Then in each time window, for each channel (graph node), the orientation-independent total intensity (TOT) is summed.

Step S8: For each patient's SEEG channel, we summarize the mean value of the total intensity 8s before the onset and the HFEI, and normalize them respectively. Combining the TOT and HFEI, we will get what is called the connectivity high-frequency epileptogenicity Composite Index of Indexes (cHFEI):

The area under the ROC curve (AUC) of different epileptogenic zone localization methods was calculated to evaluate the performance of the present invention.

The comparison results of the area under the ROC curve (AUC) of different methods are shown in Figure 2. SEEG data of 20 epileptic patients with different seizure patterns were analyzed in the present invention. Figure 2(A) is a comparison of the overall AUC values of different methods in all patients. Compared with HFO-based detection methods only (EI, HFEI), the average AUC value of cHFEI was significantly greater than that of HFEI and EI, and the prediction effect was the best. At the same time, the variances of HFEI and EI are both large, while the variance of cHFEI is relatively small, indicating that the present invention has high stability and is applicable to different types of epileptic seizures. Figure 2(B) is the comparison of AUC values of different methods in each patient. In the present invention, the AUC values of patients s1 and s19 are slightly lower than HFEI, and the AUC values of other patients are higher than those of the other two methods. In particular, the AUCs of patients s9, s16, and s20 were all 1, which was highly consistent with the epileptogenic zone marked by clinicians.

The above results show that the present invention can accurately locate the epileptogenic zone of patients with different epileptic seizure patterns, and has potential application value in epilepsy treatment.

The above-mentioned embodiments are only descriptions of preferred implementations of the present invention, and are not intended to limit the scope of the present invention. All such modifications and improvements should fall within the scope of protection defined by the claims of the present invention.

Claims (10)

1. an epileptogenic zone localization method based on high-frequency oscillation and connectivity, is characterized in that comprising the following steps: Step S1: collecting SEEG data and selecting representative channels; Step S2: Filter and acquire 54-200Hz signal and select baseline data and target data; Step S3: Obtain normalized high frequency energy NHFE; Step S4: Calculate the time coefficient TC and the energy coefficient EC according to NHFE, and obtain the high-frequency epileptogenicity index HFEI; Step S5: filtering to obtain a 12-45Hz signal; Step S6: calculating nonlinear regression analysis; Step S7: Calculate the total intensity TOT; Step S8: Define and calculate the connectivity high-frequency epileptogenicity index cHFEI, and locate the epileptogenic zone according to the cHFEI. 2. A method for locating the epileptogenic zone based on high-frequency oscillation and connectivity according to claim 1, characterized in that: in the step S1, according to the method of stereotaxy, twenty subjects with different epileptic seizure patterns were collected. According to the patient's three-dimensional EEG SEEG data, 10-30 representative channels are selected to cover the main area of the epileptic seizure. 3. A method for locating the epileptogenic zone based on high-frequency oscillation and connectivity according to claim 1, characterized in that: in the step S2, a second-order IIR notch digital filter and a fifth-order IIR Butterwater The original signal was filtered to the high-frequency band of 54-200Hz by the Sri Lankan digital filter; the baseline data and target data were manually selected. The principle of baseline data selection was SEEG data without obvious abnormal signals before the seizure, and the target data had to cover the entire seizure process. 4. A method for locating an epileptogenic zone based on high frequency oscillation and connectivity according to claim 1, characterized in that: in the step S3, the bandpass signal is converted into a high Frequency energy spectrum, and calculate the average value of the high-frequency energy of the baseline data, as the baseline value of high-frequency energy; divide the high-frequency energy of all channels by the high-frequency energy baseline value of this channel to obtain the normalized high-frequency energy NHFE . 5. A method for locating the epileptogenic zone based on high-frequency oscillation and connectivity according to claim 1, characterized in that: the specific operation of the step S4 is: Calculate the time-to-onset threshold for each channel i, defined as the maximum value of the baseline NHFE plus ten times the standard deviation of the baseline NHFE: threshold _onsettime = max(NHFE _BL )+10σ(NHFE _BL ), When the normalized energy of the target data of each channel exceeds the threshold, this moment is judged as the time when the abnormal activity starts, and the channels are sorted according to the starting time of each channel, and the time coefficient TC is defined as the reciprocal of the order of each channel ; Use NHFE to calculate the average energy within 250ms before and after the earliest start of each channel as the energy coefficient EC, and finally get the HFEI of each channel i:

6. A method for locating the epileptogenic zone based on high-frequency oscillation and connectivity according to claim 1, characterized in that: in the step S5, the original data is band-pass filtered, and the passband is the β-γ frequency band , 12-45Hz, then downsampled to 256Hz. 7. A kind of epileptogenic zone localization method based on high-frequency oscillation and connectivity according to claim 1, is characterized in that: in described step S6, calculate nonlinear regression analysis based on h ² nonlinear correlation index, in each Performs a piecewise linear regression between the signals, testing all shifts of one signal relative to the other within a maximum lag; for two signals x and y, the piecewise linear approximation f of the transfer function between x and y is defined as : y _n-τ = f(x _n )+e _n ,

^h2 measures the goodness-of-fit of the regression, explained by approximating f:

8. A kind of epileptogenic zone localization method based on high-frequency oscillation and connectivity according to claim 1, is characterized in that: in described step S7, calculates all paired h ² values, use length is 3s, step Sliding windows of length 1 s with a maximum delay of 0.1 s yield a binary connectivity matrix; then in each time window, for each channel, the orientation-independent total intensity TOT is summed. 9. A kind of epileptogenic zone positioning method based on high-frequency oscillation and connectivity according to claim 1, characterized in that: in the step S8, for each patient's SEEG channel, sum up the total intensity of the 8s before the onset and HFEI were normalized respectively, combined with TOT and HFEI, the combined index cHFEI of connectivity high-frequency epileptogenicity index was obtained:

10. A method for locating the epileptogenic zone based on high-frequency oscillation and connectivity according to claim 1, characterized in that: the positioning of the epileptogenic zone is accurate and stable.

Images