In vivo animation of auditory-language-induced gamma-oscillations in children with intractable focal epilepsy

Abstract

We determined if high-frequency gamma-oscillations (50- to 150-Hz) were induced by simple auditory communication over the language network areas in children with focal epilepsy. Four children (aged 7, 9, 10 and 16 years) with intractable left-hemispheric focal epilepsy underwent extraoperative electrocorticography (ECoG) as well as language mapping using neurostimulation and auditory-language-induced gamma-oscillations on ECoG. The audible communication was recorded concurrently and integrated with ECoG recording to allow for accurate time lock on ECoG analysis. In three children, who successfully completed the auditory-language task, high-frequency gamma-augmentation sequentially involved: i) the posterior superior temporal gyrus when listening to the question, ii) the posterior lateral temporal region and the posterior frontal region in the time interval between question completion and the patient's vocalization, and iii) the pre- and post-central gyri immediately preceding and during the patient's vocalization. The youngest child, with attention deficits, failed to cooperate during the auditory-language task, and high-frequency gamma-augmentation was noted only in the posterior superior temporal gyrus when audible questions were given. The size of language areas suggested by statistically significant high-frequency gamma-augmentation was larger than that defined by neurostimulation. The present method can provide in vivo imaging of electrophysiological activities over the language network areas during language processes. Further studies are warranted to determine whether recording of language-induced gamma-oscillations can supplement language mapping using neurostimulation in presurgical evaluation of children with focal epilepsy.

Figure 1. Simultaneous Recording of ECoG and Vocal Sound Waves in Patient 1

(A) An example of ECoG trace suitable for quantitative analysis is shown with a lowfrequency filter of 53-Hz and a high-frequency filter of 300-Hz. Vocal sound waves were simultaneously recorded with intracranial ECoG. The time-lock trigger was placed at the onset of patient’s vocalization. (B) Vocal sound wave on Cool Edit Pro Software is shown, and this was used to visually and audibly aid in the manual determination of the onset of the patient’s vocalization.

Figure 2. Language Mapping in Patient 1

(A) Neurostimulation: Stimulation of an electrode pair of the posterior temporal lobe (Ch33 & 34; denoted by a yellow box) induced language comprehension deficit. Stimulation of an electrode pair of the posterior frontal lobe (Ch54 & 70; denoted by a pink-blue box) resulted in pure speech arrest at 6 mA of stimulation and induced speech arrest associated with facial movement at 9 mA of stimulation. Stimulation of three electrode pairs on the post- and pre-central gyri (Ch53 & 54; 52 & 53; 67 & 68; denoted by green boxes) resulted in tingling of the mouth including the tongue. Stimulation of electrode pairs (Ch75 & 76; 83 & 84; denoted by green boxes) resulted in sensory responses involving the right upper extremity. Language function was not satisfactorily assessed in the following electrode sites, where neurostimulation induced prominent motor responses and stimulation had to be prematurely terminated before completion of a question-and-answer trial. Facial movement was induced by stimulation of an electrode pair of the inferior portion of the pre-central gyrus (Ch69 & 77; denoted by a blue box), and motor responses involving the right upper extremity were induced by stimulation of electrode pairs of the superior portion of the pre- and post-central gyri (Ch83 & 91; 91 & 92; 90 & 97; 98 & 99; denoted by blue boxes). Surgical resection of the superior precentral gyrus and the superior premotor region resulted in worsening of hemiparesis in the right-sided upper extremity but no postoperative language deficits were noted. (B - G) ECoG time-frequency analysis time-locked to patient’s vocalization: This analytic method was designed to evaluate sequential brain activation associated with comprehension, word retrieval, and vocalization. A greater than 50% increase in ’gamma-range amplitude’ (across 50- to 150-Hz frequency bands) was noted in the left superior temporal gyrus at 1350-msec prior to the patient’s vocalization (B), in the left middle temporal gyrus at 590-msec prior to the patient’s vocalization (C), in the left middle temporal gyrus, inferior frontal gyrus and superior frontal gyrus at 430-msec prior to the patient’s vocalization (D), in the left superior frontal gyrus as well as pre- and postcentral gyri at the onset of patient’s vocalization (E), and in the left superior temporal gyrus, superior frontal gyrus as well as pre- and post-central gyri at 130-msec after the patient’s vocalization (F). (G) Sound waveform recording data showed no temporal overlap between the sound of auditory questions and that of patient’s vocalization. Highfrequency gamma-augmentation began to involve the superior temporal gyrus (Ch 36) at 1,750-msec prior to the onset of patient’s vocalization, the middle temporal gyrus (Ch 33) at 740-msec prior to the onset of patient’s vocalization, the inferior frontal gyrus (Ch 56) at 510-msec prior to the onset of patient’s vocalization, the inferior pre-central gyrus (Ch 54) at 470-msec prior to the patient’s vocalization, and the superior temporal gyrus (Ch 36) at 70-msec after the onset of patient’s vocalization. (H) ECoG time-frequency analysis time-locked to auditory questions: This analytic method was designed to evaluate brain activation associated with the initiation of auditory question. Sound waveform recording data showed the evidence of temporal overlap between the sounds derived from auditory questions and patient’s vocalization. Thus, high-frequency gamma augmentation in the superior temporal gyrus (Ch 36) was probably induced by both auditory questions and patient’s vocalization at some point. The time-frequency matrixes for the entire subdural electrode sites are presented as supplementary data on the website.

Figure 2. Language Mapping in Patient 1

(A) Neurostimulation: Stimulation of an electrode pair of the posterior temporal lobe (Ch33 & 34; denoted by a yellow box) induced language comprehension deficit. Stimulation of an electrode pair of the posterior frontal lobe (Ch54 & 70; denoted by a pink-blue box) resulted in pure speech arrest at 6 mA of stimulation and induced speech arrest associated with facial movement at 9 mA of stimulation. Stimulation of three electrode pairs on the post- and pre-central gyri (Ch53 & 54; 52 & 53; 67 & 68; denoted by green boxes) resulted in tingling of the mouth including the tongue. Stimulation of electrode pairs (Ch75 & 76; 83 & 84; denoted by green boxes) resulted in sensory responses involving the right upper extremity. Language function was not satisfactorily assessed in the following electrode sites, where neurostimulation induced prominent motor responses and stimulation had to be prematurely terminated before completion of a question-and-answer trial. Facial movement was induced by stimulation of an electrode pair of the inferior portion of the pre-central gyrus (Ch69 & 77; denoted by a blue box), and motor responses involving the right upper extremity were induced by stimulation of electrode pairs of the superior portion of the pre- and post-central gyri (Ch83 & 91; 91 & 92; 90 & 97; 98 & 99; denoted by blue boxes). Surgical resection of the superior precentral gyrus and the superior premotor region resulted in worsening of hemiparesis in the right-sided upper extremity but no postoperative language deficits were noted. (B - G) ECoG time-frequency analysis time-locked to patient’s vocalization: This analytic method was designed to evaluate sequential brain activation associated with comprehension, word retrieval, and vocalization. A greater than 50% increase in ’gamma-range amplitude’ (across 50- to 150-Hz frequency bands) was noted in the left superior temporal gyrus at 1350-msec prior to the patient’s vocalization (B), in the left middle temporal gyrus at 590-msec prior to the patient’s vocalization (C), in the left middle temporal gyrus, inferior frontal gyrus and superior frontal gyrus at 430-msec prior to the patient’s vocalization (D), in the left superior frontal gyrus as well as pre- and postcentral gyri at the onset of patient’s vocalization (E), and in the left superior temporal gyrus, superior frontal gyrus as well as pre- and post-central gyri at 130-msec after the patient’s vocalization (F). (G) Sound waveform recording data showed no temporal overlap between the sound of auditory questions and that of patient’s vocalization. Highfrequency gamma-augmentation began to involve the superior temporal gyrus (Ch 36) at 1,750-msec prior to the onset of patient’s vocalization, the middle temporal gyrus (Ch 33) at 740-msec prior to the onset of patient’s vocalization, the inferior frontal gyrus (Ch 56) at 510-msec prior to the onset of patient’s vocalization, the inferior pre-central gyrus (Ch 54) at 470-msec prior to the patient’s vocalization, and the superior temporal gyrus (Ch 36) at 70-msec after the onset of patient’s vocalization. (H) ECoG time-frequency analysis time-locked to auditory questions: This analytic method was designed to evaluate brain activation associated with the initiation of auditory question. Sound waveform recording data showed the evidence of temporal overlap between the sounds derived from auditory questions and patient’s vocalization. Thus, high-frequency gamma augmentation in the superior temporal gyrus (Ch 36) was probably induced by both auditory questions and patient’s vocalization at some point. The time-frequency matrixes for the entire subdural electrode sites are presented as supplementary data on the website.

Figure 3. Language Mapping in Patient 3

(A) Neurostimulation: Stimulation of two electrodes pairs on the occipital lobe resulted in visual symptoms (Ch1 & 2; 49 & 50; denoted by light-blue boxes). Stimulation of an electrode pair of the inferior pre-central gyrus (Ch105 &113; denoted by a pink-blue box) induced speech arrest associated with throat movement. Stimulation of an electrode pair of the inferior post-central gyrus (Ch106 & 114; denoted by a green box) resulted in tingling of teeth. Language function was not satisfactorily assessed in the following electrode sites, where neurostimulation induced positive motor responses and stimulation had to be prematurely terminated before completion of a question-and-answer trial. Stimulation of an electrode pair of the pre- and post-central gyrus (Ch121 & 129; denoted by a blue box) resulted in movement of mouth. Stimulation of an electrode pair of the post-central gyrus (Ch122 & 129; denoted by a blue box) resulted in movement of the thumb. Stimulation of a pair of the medial frontal region (Ch6 & 7; denoted by a blue box) resulted in tonic extension of the bilateral upper extremities. (B) ECoG time-frequency analysis time-locked to patient’s vocalization: This analytic method was designed to evaluate sequential brain activation associated with comprehension, word retrieval, and vocalization. No cortical activation represented as gamma-augmentation was observed in the superior temporal gyrus (Ch 93) during auditory questions. High-frequency gamma-augmentation began to involve the posterior inferior temporal gyrus (Ch 55) at 1,220-msec prior to the onset of patient’s vocalization, the inferior frontal gyrus (Ch 103) at 590-msec prior to the onset of patient’s vocalization, and the inferior pre-central gyrus (Ch 129) immediately prior to the patient’s vocalization. (C) ECoG time-frequency analysis time-locked to auditory questions: This analytic method was designed to evaluate brain activation associated with the initiation of auditory question. High-frequency gamma augmentation began to involve the left superior temporal gyrus (Ch 93) at 70-msec after the onset of auditory questions, the posterior inferior temporal gyrus (Ch 55) at 1,020-msec after the onset of auditory questions, and the inferior frontal gyrus (Ch 103) at 910-msec after the onset of auditory questions. The time-frequency matrixes for the entire subdural electrode sites are presented as supplementary data on the website.