Subtemporal Multiple Hippocampal Transection with/without CA1-Subiculum Disconnection for Medically Intractable Temporal Lobe Epilepsy

Abstract

Transsylvian selective amygdalohippocampectomy resulted in postoperative verbal memory decline in patients with mesial temporal lobe epilepsy of the language-dominant side. Mapping whole-brain connectivity changes have been studied recently of different surgical resection approaches for temporal lobe epilepsy. The subtemporal resection is the least disruptive to long-range connectivity, which may explain its better cognitive outcome. Finally, the authors introduced subtemporal multiple hippocampal transections technique in a case of hippocampal sclerosis negative left mesial temporal lobe epilepsy, and postoperative neuropsychological examinations revealed improvement of cognitive function immediately after the operation contrasting transsylvian multiple hippocampal transections in which verbal memory remains dropped. The authors introduced another new operation to left mesial temporal lobe epilepsy patient with hippocampal sclerosis by multiple hippocampal transections plus disconnection between CA1 and subiculum at the hippocampal head. Operative result is satisfactory in terms of neuropsychological and operative outcome.

1. Introduction

The limbic lobe is located at the most medial portion of the cerebral hemisphere, like as a band surrounding the orifice into the lateral ventricle. The limbic lobe is mainly consisted of the hippocampal formation, the amygdaloid complex, and the cingulate cortices. These areas are concerned with the basic brain higher functions, such as emotion, memory, attention, cognition, and so on. Each area has the specific structural organization and the fiber connections executing the specific function. The circuit of memory, so-called Papez circuit, is consisted of the hippocampal formation, mammillary body, anterior and midline thalamic nuclei, posterior cingulate cortex, and the retrohippocampal cortices. On the other hand, the circuit of the emotion, so-called Yakovlev circuit, is consisted of the amygdala, the mediodorsal thalamic nucleus, anterior cingulate cortex, and the orbitofrontal cortex. Although memory and emotion are processed on the independent circuit, there are also several structures where the fibers from the two circuits meet together, such as the nucleus accumbens, the entorhinal cortex, and the hypothalamic area. Memory information is stored efficiently only when the brain is in appropriate state for memory acquisition. This state, the motivating state for memory, is determined by the levels of awareness, cognition, attention, emotion, and other influences. The motivating state might change signal processing in the hippocampus. Classical anatomical observations with Golgi staining of hippocampal neurons are the basis for the simple trisynaptic circuit (dentate granule cells (DG)-Cornu Ammonis 3 (CA3), CA3 to CA1 concept of hippocampal function (Figure 1).

Figure 1.

Schema illustrating disconnection between CA1 and subiculum (red line). Yellow arrow is mossy fiber, blue arrow is Schaffer collateral, green arrow is alveus hippocampi, and red line is disconnection between CA1 and subiculum just until hippocampal sulcus. This histology is sectioned from normal hippocampus taken from cadaver without central nervous lesion.

Recent anatomical work has revealed much richer synaptic connections between hippocampal neuron subfields (DG, CA3, CA2, and CA1) and wide distribution of axons along the longitudinal axis of the hippocampus [1].

Temporal lobe epilepsy (TLE) is involving the limbic system, especially amygdala and hippocampus which influences emotion, memory, attention, and cognition.

Despite various medical drugs have been tried to control intractable TLE, surgical treatment gives better seizure control comparing to prolonged medical treatment [2]. But in this randomized control study, substantial number (around 5%) of surgical patients complained of postoperative memory difficulty, but the authors acknowledge that although this complication is important, but the benefit is worth the risk. Concerning such memory problems, in surgical population, MR negative language-dominant-side mesial TLE group, so-called paradoxical temporal lobe epilepsy (PTLE) [3], is at great risk of postoperative memory deficits. Until now, neurosurgeons have made great efforts to stop such a postoperative memory decline by adopting various operative approaches, such as reducing size of temporal lobectomy, selective amygdalohippocampectomy, and recent radiosurgical interventions [4].

In this chapter of this book, our further efforts to escape from such a postoperative memory decline such as subtemporal multiple hippocampal transection (MHT) with or without disconnection between cornu ammonis (CA1) and subiculum will be presented in detail (Figure 1). These techniques and results are never reported in the world literature.

2. Neuropsychological results of various surgical treatments for TLE

Temporal lobe epilepsy (TLE) is one of the most intractable epilepsy involving the limbic system, especially amygdala and hippocampus which influences emotion, memory, attention, and cognition.

Despite various medical drugs have been tried to control intractable TLE, surgical treatment gives better seizure control comparing to prolonged medical treatment [2]. Concerning memory problems, in surgical population, MR negative language-dominant-side mesial TLE group, so-called paradoxical temporal lobe epilepsy (PTLE) [3], is at great risk of postoperative memory deficits. Until now, neurosurgeons have made great efforts to stop such a postoperative memory decline by adopting various operative approaches, such as reducing size of temporal lobectomy, selective amygdalohippocampectomy, and recent radiosurgical interventions [4].

Morino et al. reported comparison of neuropsychological results after selective amygdalohippocampectomy versus anterior temporal lobectomy (ATL) [5]. Transsylvian selective amygdalohippocampectomy (TSSAH) is an operative technique planned to spare unaffected brain region during surgical treatment for mesial temporal lobe epilepsy (MTLE). In contrast to standard anterior temporal lobectomy (ATL), the advantages of TSSAH with respect to postoperative cognitive outcome are equivocal without randomized control study. Morino et al. compared cognitive function before and after surgery in 49 patients with unilateral mesial temporal lobe seizures who underwent either ATL (n = 17) or TSSAH (n = 32). All patients received neuropsychological testing before and 1 year after surgery. The intelligence quotient (IQ) increased postoperatively in both surgical groups. Memory evaluation in the ATL group revealed a postoperative decline in nonverbal memory after right-sided resection and a postoperative decline in verbal memory after left-sided resection. In the TSSAH group, there was a slight postoperative decline only in verbal memory after left-sided resection, but other memory function was preserved. There was significant postoperative improvement in verbal memory after right-sided resection. Overall, memory function was better preserved in the TSSAH group than in the ATL group.

There is another change of operative procedure invented by Shimizu et al. [6], that is, transsylvian multiple hippocampal transection (TSMHT).

Table 1 is a summary of memory scores by transsylvian multiple hippocampal transection (TSMHT) for PTLE [6, 7]. Preoperative verbal memory score dropped 1 month after surgery, but 1 year after surgery improved up to preoperative level.

	Lt (n = 15)						Rt (n = 11)
	Preop.	1MP	1YP	Preop.-1MP	Preop.-1YP	1MP-1YP	Preop.	1MP	1YP	Preop.-1MP	Preop.-1YP	1MP-1YP
Verbal	75.7 (4.8)	73.2 (4.9)	75.0 (6.3)	p = 0.3257	p = 0.8754	p = 0.7069	77.3 (5.5)	81.3 (6.0)	83.8 (6.1)	p = 0.1887	p = 0.0970	p = 0.5225
Performance	77.7 (6.3)	81.4 (6.3)	79.6 (7.5)	p = 0.2547	p = 0.7187	p = 0.7293	77.1 (7.1)	84.4 (7.2)	91.8 (7.8)	p = 0.0794	†p = 0.0071	p = 0.1340
Total	74.0 (5.8)	74.3 (5.9)	73.9 (6.8)	p = 0.8922	p = 0.9875	p = 0.9214	74.4 (6.6)	79.4 (6.7)	87.3 (7.2)	p = 0.1262	†p = 0.0033	p = 0.0603
No significant improvement or decline are observed for verbal intelligent quotient (IQ), performance IQ, and total IQ.
	Verbal memory (WMS-R) Morino et al. [7]
	Preop.	1MP	1YP	Preop.-1MP	Preop.-1YP	1MP-1YP	Preop.	1MP	1YP	Preop.-1MP	Preop.-1YP	1MP-1YP
	86.1 (5.9)	71.3 (6.2)	79.8 (6.0)	*p = 0.04	p = 0.1845	p = 0.0966	89.1 (6.2)	90.8 (6.2)	97.3 (6.3)	p = 0.7165	p = 0.1009	p = 0.1922

Table 1.

Memory outcomes depending on the operative side by transsylvian hippocampal transection for PTLE by Morino et al. [7].

1MP: 1 month after op. 1YP: 1 year after op.

Only WMS-R verbal memory is deteriorated 1 month after operation, but it recovered to preoperative level 1 year after the operation, but never statistically improved comparing to preoperative level.

How about the neuropsychological results after subtemporal amygdalohippocampectomy (sSAH) for TLE [8, 9]. Figure 2 shows combined neuropsychological results of sSAH by Takaya (left panel) [8] and Hori (right panel) [9]. Both WMS-R and WAIS-R scores show significant improvements postoperatively comparing to preoperative levels. This is the difference between TSSAH and sSAH.

Figure 2.

Left panel adopted from Takaya et al. [8], and right panel from Hori et al. [9]. Left panel: cognitive improvement might result from a combined effect of good seizure control and minimize the regions of the brain with postoperative functional impairment. Improved cerebral function in terms of WMS-R scores in mesial temporal lobe epilepsy after sSAH was demonstrated [8]. Right panel: bar graphs showing changes in IQs in patients in whom the language-dominant and nondominant hemisphere was resected by sSAH [9].

Usami et al. reported their operative results by transsylvian approach with multiple hippocampal transection and multiple subpial transection (MST) with lesionectomy (TSMHT +MST/L) technique [10] for TLE. As it is clearly shown, follow-up results show some deterioration of scores, especially VIQ. Figures 2 and 3 show the difference of postoperative cognitive function between TSSAH MHT + MST/L and sSAH [8, 9, 10].

Figure 3.

Adopted from Usami et al. [10].

There were changes in neuropsychological function after surgery on the verbally dominant side (n = 12). There were no significant differences between the preoperative indices and those at the last visit. The values were as follows (mean ± SD preoperatively, at the last visit; p-value): verbal memory (85 ± 13, 78 ± 18; 0.14), visual memory (94 ± 24, 102 ± 16; 0.08), general memory (85 ± 16, 82 ± 19; 0.29), delayed recall (79 ± 18, 87 ± 19; 0.09), attention and concentration (89 ± 21, 88 ± 22; 0.10). Regarding IQ, there was a significant difference in VIQ (87 ± 19, 80 ± 19; 0.045*), but not in PIQ (89 ± 28, 88 ± 23; 0.42) or FIQ (86 ± 25, 82 ± 21; 0.16) [10].

3. Subtemporal amygdalohippocampectomy

There are many surgical techniques to cure medically intractable TLE such as conventional temporal lobectomy. But, concerning the language-dominant-side TLE without hippocampal sclerosis (PTLE) surgical removal of mesial temporal structures may result in neuropsychological problems, especially decline of verbal memory. Usually, temporal lobectomy 4–4.5 cm away from the temporal tip is used for language-dominant-side TLE [2]. Postoperative verbal memory decline is a major concern especially for language-dominant-side TLE without HS so-called PTLE [3, 7]. For mesial TLE of which amygdala and hippocampal head are epileptic foci, selective amygdalohippocampectomy is used by various routes including transsylvian, through superior temporal gyrus or sulcus (T1), middle temporal gyrus or sulcus (T2), and inferior temporal gyrus or sulcus (T3) depending on the preference of neurosurgeons (Figure 4). In 1993, subtemporal amygdalohippocampectomy (sSAH) technique has been introduced for mesial temporal lobe epilepsy to abolish postoperative neuropsychological deterioration observed in usual anterior temporal lobectomy [11]. The conventional subtemporal approach has been modified to diminish temporal lobe compression pressure and the risk of damage to the temporal stem. In this technique, the approach has been changed from usual anterolateral approach to posterolateral, thereby avoiding the voluminous and deeply embedded anterior temporal lobe in the middle fossa. By this approach, the retraction pressure is decreased and the temporal stem which is important bottleneck of temporal information were spared. To date, the authors using this approach have operated on more than 50 patients with medically intractable temporal lobe epilepsy whose epileptic foci were in the mesial temporal lobe structure; the inferior temporal gyrus, the temporal tip, the vein of Labbe, and the ventral bridging veins were preserved using with dissecting technique without adverse events. It can be used to remove as much of the posterior hippocampus as necessary, and it can be extended to conventional lobectomy if necessary. If there is some drawback in this approach, a part of basal temporal language area is sacrificed in order to reach temporal horn (Figure 5A–C). One patient among 50 patients with HS negative language-dominant-side TLE (PTLE) has shown postoperative severe memory deficits. Although the patient has shown gradual improvement of her memory in these 20 years during her postoperative period, she is still complaining of memory problem and it is compromising her daily job, even if operated upon by sSAH.

Figure 4.

A,B,C Microscopic procedure of subtemporal amygdalohippocampectomy. A: Temporal base is retracted, and fusiform gyrus is identified (double arrows). A small bridging vein was just riding on the collateral sulcus, and the vein is preserved. B: Fusiform gyrus, collateral sulcus is identified, and temporal horn is opened, then amygdala and hippocampal head are identified. At the choroid plexus (arrow head), the anterior choroid artery (arrow) is identified at the tips of the forceps. The suction tube touches the surface of the hippocampus. After the hippocampus (about 2.5 ∼ 3.0 cm usually in length) anterior to the choroidal point is removed, the amygdala, the parahippocampal gyrus, and the uncus were removed, so that in every patient the amygdala can be removed en bloc, and neuropathologically is examined. It is different point between TSSAH and sSAH. C: After sSAH, PCA (black star), basal vein (small black star), optic nerve (short arrow), and anterior choroidal artery (long arrow) are well exposed.

Figure 5.

Types and Frequency of HS in our operative specimen in terms of ILAE Classification. HS type 1 (25 of 41 cases, 61%) is equivalent to ‘classical’ Ammon’s horn sclerosis in which neuronal loss and gliosis is most severe in CA1, followed by CA3, CA4, with relative sparing of CA2 and often associated with loss of dentate granule cells and/or dispersion. HS type 2 represents neuronal loss and gliosis almost confined to CA1 (CA1 sclerosis), and only 1 case (2%) was identified in our study. HS type 3 (7 cases, 17%) is characterized by a reverse distribution of the sclerotic lesion to HS type 1, in which neuronal loss and gliosis is the most severe in CA4 followed by CA3, with relative sparing of CA2 and CA1, that is equivalent to endofolium sclerosis (EFS). In addition to these three HS types, we also identified 8 cases (19%) without apparent neuronal loss and gliosis (no HS).

4. Minimally invasive subtemporal approach

Recently, Busby et al. [12] reported whole-brain tract changes after mapping and analyzed the potential impact of different surgical resection approaches for TLE. The main aim of this study was to perform systematic “pseudo-neurosurgery” based on existing resection methods on healthy neuroimaging data and measuring the effect on long-tract connectivity. They use anatomical connectivity mapping (ACM) to determine long-range disconnection, which is complementary to existing measures of local integrity such as fractional anisotropy or mean diffusivity. ACMs were generated for each diffusion scan in order to compare whole-brain connectivity with an “ideal resection,” nine anterior temporal lobectomy and three selective approaches. For en bloc resections, as distance from the temporal pole increased, reduction in connectivity was evident within the arcuate fasciculus, inferior longitudinal fasciculus, inferior front-occipital fasciculus, and the uncinate fasciculus.

Increasing the height of resections dorsally reduced connectivity within the uncinate fasciculus. sSAH was associated with connectivity modes most similar to the “ideal” baseline resection, compared to TSSAH and middle-temporal approaches. In conclusion, Busby N, et al. showed the utility of ACM in assessing long-range disconnections/disruptions during temporal lobe resections, where they identified the sSAH as the least disruptive to long-range connectivity which may explain its better cognitive outcome. Of course, magnetic resonance (MR)-guided focused ultrasound treatment of mesial TLE is an ideal treatment if properly sonicated at the key structures of intractable TLE.

In 2021, Whiting AC and Smith KA, et al. [13] reported seizure and neuropsychological outcomes in a large series of selective amygdalohippocampectomies with a minimally invasive sSAH which is almost similar to our approach.

All patients in this study had at least 1 year of follow-up (mean [SD] 4.52 [2.57] years), of whom 57.9% (88/152) had Engel Class I seizure outcomes.

Engel’s classification is as follows:

Class I: Seizure free or no more than a few early, nondisabling seizures; or seizures upon drug withdrawal only.

Class II: Disabling seizures occur rarely during a period of at least 2 years; disabling seizures may have been more frequent soon after surgery; nocturnal seizure.

Class III: Worthwhile improvement; seizure reduction for prolonged periods but less than 2 years.

Class IV: No worthwhile improvement; some reduction, no reduction, or worsening are possible.

The patients with at least 2 years of follow-up (mean [SD] 5.2 [2.36] years), 56.5% (70/124) had Engel Class I seizure outcomes. Of the 152 patients with at least 1 year of clinical follow-up, only 38 (25%) completed both preoperative and postoperative neuropsychological testing by a neuropsychologist. The mean difference in scores was statistically significant in the dominant hemisphere group for the RAVLT–short delay test (p = 0.02) and the BNT (Boston Naming Test) (p = 0.04). The mean difference in scores was statistically significant in the nondominant group only with the BNT (p = 0.04). Many patients were unable to complete both preoperative and postoperative neuropsychological examinations for a variety of reasons. Concerning this decline, it might be influenced by some bias, that is, only 25% of all patients are examined, and a small number of neuropsychologically declined group of patients are prone to be examined.

Adverse events were low, with a 1.3% (2/152) permanent morbidity rate and 0.0% mortality rate.

This study reports a large series of patients who have undergone sSAH, with a minimally invasive technique. The sSAH approach described in this study appears to be a safe, effective, minimally invasive technique for the treatment of MTLE among the surgical methods ever reported in the literature.

5. Neuropsychological results after subtemporal amygdalohippocampectomy

In 2007, the authors [9] evaluated operative, neuropathological, and neuropsychological results after selective subtemporal amygdalohippocampectomy (sSAH) for refractory temporal lobe epilepsy in patients who were observed for at least 2 years after surgery. More than 26 consecutive patients underwent sSAH for non-lesional, medically refractory TLE. Neuropsychological evaluation using the Wechsler Adult Intelligence Scale (WAIS) was done before surgery in all patients, 2 months after surgery in 24 patients, and at 2-year follow-up in 19 patients. The data were compared between the 13 patients in whom the language-dominant hemisphere was surgically treated and the six patients in whom the language-nondominant hemisphere was treated. After surgery, 84% of the patients showed either Engel Class I or II seizure outcome. There were no permanent subjective complications except postoperative memory impairment in one patient with normal intelligence without HS. Neuropathological examination revealed HS in 19 patients. No significant differences in IQ and verbal memory test scores were observed between the patients in whom the language-dominant hemisphere was treated and those in whom the language-nondominant hemisphere was treated. One patient without HS whose language-dominant hemisphere was treated by sSAH, postoperative memory loss is only her complication. In this patient, although her memory has slightly improved, her job is compromised by memory loss. Considering this situation, neurosurgeons should explore better operative technique to escape from such postoperative memory loss, especially for language-dominant-side TLE without HS (PTLE).

Significant postoperative improvements in verbal IQ, performance IQ (PIQ), and full-scale IQ (FIQ) were observed over time. No significant differences were found between pre- and postoperative verbal memory test scores, and no subjective visual field loss was marked in any patient. Thus, sSAH provides good surgical and neuropsychological results and does not cause significant postoperative decline of verbal memory even if performed on the language-dominant side. In Figure 2, graphs depicting changes in IQs in patients in whom the language-dominant hemisphere was resected and patients in whom the language-nondominant hemisphere was resected, changes were time-dependent (VIQ, P = 0.0107; PIQ, p = 0.0002; FIQ, p = 0.0003), with no significant differences between the dominant and nondominant hemisphere groups (VIQ, p = 0.9102, PIQ, p = 0.7454; FIQ, p = 0.8361), and significant increases in VIQ, PIQ, and FIQ were observed over time.

Takaya et al. [8] evaluated the effects of sSAH on cerebral glucose metabolism and memory function in 15 patients with medically intractable MTLE with HS using [18F]-fluorodeoxyglucose PET (FDG-PET) and the Wechsler Memory Scale-Revised (WMS-R). The patients were evaluated before and 1–5 years (mean 2.6 years) after surgery. In patients with MTLE of the language-dominant hemisphere, the basal temporal language area was preserved by this surgical approach. Postoperative glucose metabolism increased in extratemporal areas ipsilateral to the affected side, such as the dorsolateral prefrontal cortex, and the dorsomedial and ventromedial frontal cortices. Glucose metabolism also increased in the bilateral inferior parietal lobules and in the remaining temporal lobe regions remote from the resected mesial temporal region, such as the superior temporal gyrus and the temporal pole. By contrast, postoperative glucose metabolism decreased only in the mesial temporal area near the resected region. Postoperative verbal memory, delayed recall, and attention/concentration scores were significantly better than preoperative scores regardless of the resected side (Figure 2 Left). This study suggests that the selective removal of the epileptogenic region in MTLE using subtemporal approach improved cerebral glucose metabolism in the areas receiving projections from the affected mesial temporal lobe. Cognitive improvement might result from a combination of good seizure control and minimize the area of the brain with postoperative functional impairment. Improved cerebral function in terms of WMS-R scores in mesial temporal lobe epilepsy after sSAH was demonstrated in Figure 2 Left [8].

Judging from these figures, the difference between our neuropsychological results and Takaya’s results is not clear, indicating that sparing incision of basal language area in Takaya’s series may not influence the results.

6. Histological classification of hippocampal sclerosis

As demonstrated in Figure 5, HS type 1 (25 of 41 cases, 61%) is equivalent to “classical” Ammon’s horn sclerosis in which neuronal loss and gliosis are most severe in CA1, followed by CA3, and CA4, with relative sparing of CA2 and often associated with loss of dentate granule cells and/or dispersion [14, 15]. HS type 2 represents neuronal loss and gliosis almost confined to CA1 (CA1 sclerosis), and only one case (2%) was identified in our study. HS type 3 (7 cases, 17%) is characterized by a reverse distribution of the sclerotic lesion to HS type 1, in which neuronal loss and gliosis are the most severe in CA4 followed by CA3, with relative sparing of CA2 and CA1, that is equivalent to endofolium sclerosis (EFS). In addition to these three HS types, we also identified eight cases (19%) without apparent neuronal loss and gliosis (no HS). Subiculum was relatively well preserved in all cases. Granule cell dispersion is one of the abnormal structural changes that has been shown in patients with temporal lobe epilepsy. In a normal situation, the granule cells in dentate gyrus should be tightly packed. But in granule cell dispersion, the compact formation was lost, and the axons need to extend longer to reach the neighboring granule cells. It might be a consequence of a migration disorder, and the first hypothesis considers an initial injury that releases toxin(s) that affect the normal migration of granule cells. The second hypothesis concerns the role of reelin. Reelin is required for normal neuronal lamination in humans, and the lack of this expression can lead to migration defect associated with temporal lobe epilepsy.

Types of HS did not correlate with age at operation and duration of illness, suggesting that these types represent distinct pathology of MTLE, the mean age of onset in patients with type 1 sclerosis tends to be younger than those at least with no HS but this is not statistically significant (Kruskal-Wallis test), the history of initial precipitating injury is not correlated with histological subtypes or postoperative seizure control, and type 1 sclerosis seems to correlate with better postsurgical seizure outcome than other types [14, 15].

The choice of the operative procedure is important factor affecting the seizure outcome, and that lateral temporal structure is also involved in the epileptogenicity in a subset of patients with MTLE (Tables 2 and 3).

HS types	N (%)	Ammon’s horn on MRI (%)			Ammon’s horn pathology (%)			Amygdala pathology (%)					Dual pathology (%)
		n	HI-signal	Atrophy	n	N-swell	GCD	n	N-swell	DG	FG	NL & G
Type 1	25 (61.0)	25	23 (92.0)	19 (76.0)	25	24 (96.0)	24 (96.0)	21	13 (61.9)	17 (81.0)	1 (4.8)	1(4.8)	6125 (24.0)
Type 2	I (2.4)		1	0		0	1, focal		0		0	0	II 1
Type 3	7 (17.1)	7	3 (42.9)	3 (42.9)	7	1 (14.3)	4 (57.1)	6	5 (83.3)	6 (100)	0 (0.0)	0 (0.0)	3/7 (42.9)
No HS	8(19.5)	7	4 (57.1)	3 (42.9)	8	2 (25.0)	1 (12.5)	8	5 (62.5)	7 (87.5)	1 (12.5)	0 (0.0)	—
Total	41 (100)	40	31 (77.5)	25 (62.5)	41	27 (65.9)	30 (73.2)	36	23 (63.9)	31(86.1)	2 (5.6)	I (2.8)	10,133 (30.3)

Table 2.

ILAE HS type 1 is found in 25/41 (61%) patients, while amygdala sclerosis is found in only one patient (Figure 5), and dual pathology is found in 10/33 (30.3%).

Abbreviations: DG: diffuse gliosis without significant neuronal loss, FG: focal gliosis, GCD: granule cell dispersion; HI-signal: high-intensity signal on T2-weighted and/or FLAIR image, N-swell: swelling of remaining neurons; NL & G: neuronal loss and gliosis.

HS types	N (%)	Onset (y) Mean ± SD	Duration (y) Mean ± SD	IPI (%)	Engel’s class (n)
					I	II	III	IV	n/a
Type 1	25 (61.0)	12.0 ± 11.0	19.7 ± 10.1	17/25 (68.0)	19	2	4	0
Type 2	I (2.4)	29	21	0/1			1
Type 3	7 (17.1)	17.9 ± 14.0	16.1 ± 9.7	2/7 (28.6)	3	1	2	1
No HS	8 (19.5)	22.1 ± 10.0	15.4 ± 13.2	4/6 (66.7)	5	1	1	0	1
Total	41 (100)	14.7 ± 11.7	18.4 ± 10.4	22/37 (59.5)	27	4	8	1	1

Table 3.

HS types and clinical features.

Abbreviations: IPI: clinical history of initial precipitating injury; SD: standard deviation.

In 2019, Seki et al. [16] reported an analysis of proliferating neuronal progenitors and immature neurons in the human hippocampus surgically removed from control and epileptic patients. Adult neurogenesis in the mammalian hippocampus is a well-known phenomenon (Figure 6).

Figure 6.

Amygdala sclerosis. Amygdala sclerosis is detected in only one patient with HS type 1 (Table 2). Amygdala basolateral nucleus, Left: Nissl’ stain, Right: glial fibrillary acidic protein (GFAP) stain.

It remains controversial as to what extent adult neurogenesis actually occurs in hippocampus, and how brain diseases, such as epilepsy, affect human adult neurogenesis. We analyzed polysialylated neural cell adhesion molecule (PSA-NCAM) cells and proliferating neuronal progenitor (Ki67+/mammalian Hu protein B (HuB)+/ doublecortin (DCX) + cells in the surgically removed hippocampus of epileptic patients. In control patients, a substantial number of PSA-NCAM+ cells were distributed densely below the granule cell layer (GCL). In epileptic patients with granule cell dispersion, the number of PSA-NCAM+ cells were reduced, and aberrant PSA-NCAM+ cells were found. However, the numbers of Ki67+/HuB+/DCX+ cells were very low in both control and epileptic patients. The large number of PSA-NCAM+ cells and few DCX+/HuB+/Ki-67+ cells observed in the controls suggest that immature-type neurons are not recently generated neurons, and that the level of hippocampal neuronal production in adult humans is low. These results also suggest that PSA-NCAM is a useful marker for analyzing the pathology of epilepsy, but it is not evident that these bizarre PSA-NCAM neurons are the results or cause of intractable epilepsy (Figure 7). Different interpretations of the immunohistochemical results between humans and rodents should be examined in future.

Figure 7.

In control patients, a substantial number of PSA-NCAM+ cells were distributed densely below the granule cell layer (GCL). In epileptic patients with granule cell dispersion, the number of PSA-NCAM+ cells were reduced, and aberrant PSA-NCAM+ cells were found.

7. Technique of the multiple hippocampal transection (MHT)

The consequences of resection of the hippocampus, where its function is still preserved, can be a decrease in verbal memory or visual–spatial memory, intelligence, emotional and speech performance, as well as cognitive disorders. To solve this problem, Shimizu et al. [6] in 2006 reported the technique of the multiple hippocampal transection (MHT). The concept of MHT originated on the basis of multiple subpial transection in eloquent areas of the neocortex. Uda et al. [17] reported differences based on the surgical side: MHT on the nondominant side resulted in significant improvements in verbal but not visual memory, whereas MHT on the dominant side did not lead to significant increase in verbal or visual memory.

The principle of surgical treatment of mesial temporal lobe epilepsy by multiple transverse transection of the hippocampus is the mechanical disruption of the longitudinal pathways of the hippocampus.

Recent anatomical evidence suggests a functionally significant back-projection pathway from the subiculum to the CA1. A critical role for CA1-projecting subicular neurons in object-location learning and memory show that this projection modulates place-specific activity of CA1 neurons and their responses to displaced objects. Together, these experiments reveal a novel pathway by which cortical inputs, particularly those from the visual cortex, reach the hippocampal output region CA1. It is established that the hippocampus has two types of pathways: (1) trisynaptic pathways, which are located in parallel loops oriented orthogonally to the longitudinal axis of the hippocampus; (1) from entorhinal cortex to granule cell layer (perforant fiber), (2) from GCL to CA3 (mossy fiber), and (3) from CA3 to CA1 (Schaffer collateral fiber). These fiber systems are so-called trisynaptic circuit of the hippocampus (Figure 1).

There are two longitudinal pathways that run along the long axis of the hippocampus [7]. Loops of trisynaptic pathways going into the entorhinal cortex are important for processing and stabilizing memory. The longitudinal path of the hippocampus does not play an important physiological role; on the contrary, it facilitates the synchronization of pathological epileptic discharges and their propagation along the hippocampus and, further, to extrahippocampal structures, thus contributing to the development of a seizure. For the pathological electrical activity in the hippocampal neurons to develop into an epileptic seizure, synchronization of the critical number of neurons – exceeding 5 mm thickness – located in the hippocampal segment is necessary. Therefore, if longitudinal horizontal interneuronal fibers along the axis of the hippocampus are separated with an interval of 5 mm, then the pathological connection is interrupted and, an epileptic seizure stop (Figure 8).

Figure 8.

Multiple subtemporal hippocampal transection every 5 mm apart.

8. The results of long-term changes in cognitive function after MHT on the verbally dominant side

Usami et al. [10] reported the results of long-term changes in cognitive function after surgery on the verbally dominant side (n = 12). This clinical research is most recent and reliable because of containing fair number of verbally dominant side. There were no significant differences (Figure 3) between the preoperative indices and those at the last visit. Regarding intelligence, there was a significant difference in VIQ (8719,8019; 0.045*), but not in PIQ or FIQ. They concluded that, in all neuropsychological batteries, the average indices declined temporarily at 1 month, recovered to the preoperative level at 6 months, and were maintained for a long time after MHT + multiple subpial transection/lesionectomy (MST/L). There were no statistically significant differences between the preoperative and last-visit values in all batteries. In three patients, verbal memory indices dropped >20 points from preoperative figures after >5 years. Although VIQ (verbal intelligence quotient) and FIQ (full-scale intelligence quotient) declined temporarily at 1 month, they recovered to preoperative levels at 6 months. PIQ (performance intelligence quotient) and FIQ were preserved at the last visit, whereas VIQ had declined at the last visit in comparison with that on the preoperative test (p = 0.045). Judging from these results, transsylvian approach influenced this decline of VIQ comparing to our gain of VIQ for language-dominant side by sSAH operation (Figure 2 Right Panel).

The important finding was that there was a significant discrepancy between memory indices and morphologic changes of the mesial temporal lobe and associated structures.

Memory preservation: postoperative cognitive impairment has been an important and controversial issue in the surgical treatment of mTLE. Patients with PTLE have a significant risk of postoperative memory decline. Long-term observation after medial temporal resection revealed that the memory impaired by surgery did not recover over time.

Although the Wada test (sodium amytal is injected into the internal carotid artery to induce a temporary state of hemianesthesia during which language and memory function of the unaffected hemisphere are tested) is not a reliable predictor of postoperative memory decline, Usami et al. recently demonstrated that parahippocampal high-gamma activity could provide predictive information about whether the mesial temporal lobe can be resected without causing memory decline.

The postoperative decline in verbal memory impairs cognitive performance in patients with MTLE. Verbal memory function after anterior temporal lobectomy or transsylvian SAH deteriorates at the group level in patients with dominant-side MTLE, whereas it tends to improve in patients with nondominant-side MTLE (Morino et al. [5]).

In our study, an improvement in verbal memory was observed regardless of the resected side. Previous studies have reported that sSAH might escape from verbal memory decline in patients with dominant-side MTLE [7, 18, 19]. Preservation of the basal temporal language area resulted in improved verbal memory 1 year after the operation, even when the anti-epileptic drug (AED) dosage remained unchanged. Mikuni et al. also shows a long-lasting improvement in verbal memory following sSAH. The basal temporal language area is located between 10 mm and 75 mm posterior to the temporal tip and is important in processing verbal information. Verbal IQ is the ability to understand and reason using concepts framed in words, and it improved after 2 years. Performance IQ (PIQ) is designed to provide a measure of an individual’s overall visuospatial intellectual abilities score and full-scale IQ (FIQ) which is an overall score as well as scores for component abilities. Both PIQ and FIQ improved after both 2 months and 2 years postoperatively.

9. Subtemporal multiple hippocampal transection (New technique to preserve memory)

In 2021, a 51-year-old man showing left amygdala enlargement with medically intractable epilepsy patient without hippocampal sclerosis (HS) was introduced to our clinic. FDG-PET analysis (Figure 9) showed left mesial temporal lobe epilepsy with normal intelligence; in this patient, the authors adopted subtemporal selective amygdalotomy with multiple hippocampal transections (Figures 8 and 10). He has shown improved neuropsychological examination scores 3 months after surgery, and intractable seizure stopped after surgery (Engel’s Class I) and returned to his previous job immediately after the operation (Figure 11, Table 4).

Figure 9.

MR showing left amygdala enlargement, hyperintensity (FLAIR lower left and lower middle), and FDG-PET demonstrated hyper FDG activity at the amygdala (upper & lower right).

Figure 10.

After amygdalotomy multiple hippocampal transections (blue line) were performed to hippocampal head.

Figure 11.

Postoperative MRI demonstrated removal of basolateral nucleus of amygdala and preservation of hippocampus by MHT.

2021/9/27: Ope date
2021/7/1		2021/12/25
VIQ	106	VIQ	119
PIQ	109	PIQ	137
TIQ	108	TIQ	129
VCI	105	VCI	111
PRI	121	PRI	146
WMI	113	WMI	113
PSI	102	PSI	118

Table 4.

Pre- and postoperative neuropsychological examinations showing improvement in this patient.

VCI: verbal comprehension index, PRI: perceptual reasoning index, WMI: working memory index, PSI: processing speed index.

10. Pathophysiological characteristics associated with epileptogenesis in human HS

In 2017, Kitaura et al. [20] reported pathophysiological characteristics associated with epileptogenesis in human HS. Majority of seizures originate primarily from the hippocampus. They investigated epileptiform activities ex vivo using living hippocampal tissue taken from patients with MTLE. Flavoprotein fluorescence imaging and local field potential recordings revealed that epileptiform activities developed from the subiculum. Moreover, physiological and morphological experiments revealed possible impairment of K+ clearance in the subiculum affected by HS. K+ clearance is mainly regulated by astrocytes Kir 4.1 so that these findings indicate the role of astrocytes in epileptogenesis in HS. Stimulation of mossy fibers induced recurrent trans-synaptic activity in the granule cell layer of the dentate gyrus, suggesting that mossy fiber sprouting in HS also contributes to the epileptogenic mechanism presumably in addition to bizarre PSA-NCAM positive immature neurons observed in our specimen. These results indicate that pathophysiological alterations involving the subiculum and dentate gyrus could be responsible for epileptogenesis in patients with MTLE (Figure 12).

Figure 12.

Schematic representation of the results. (A) Control hippocampus showing anatomical orientation: from entorhinal cortex to granule cell layer (perforant fiber), from GCL to CA3 (mossy fiber), and from CA3 to CA1 (Schaffer collateral fiber). These fiber systems are called trisynaptic circuit of the hippocampus. Please refer to Figure 1. (B) No-HS. Enhanced activities are initiated in the subiculum and extend backward to the CA1. Please refer to Figure 8. (C) HS. Epileptogenesis in the subiculum and MFS in GCL are evident. Adopted and modified from Kitaura et al. [20].

In the No-HS group also, the activities in the subiculum and CA1 were temporarily correlated with each other. The activity in the subiculum is always being followed by that in CA1, suggesting that activity generated primarily in the subiculum was able to propagate into CA1 via feedback projection from the subiculum to CA1.

The activity generated primarily in the subiculum was able to propagate into CA1 via feedback projection from the subiculum to CA1. Kitaura et al. proposed that minimally invasive surgical approach involving disconnection of the circuit between the subiculum and the CA1 might be effective to control seizure.

11. New technique proposal for the language-dominant left TLE with HS (Type 1)

A 31-year-old woman is introduced to our clinic for the management of intractable left TLE. Her MR demonstrated typical HS. In consideration of her intractability, the amygdalotomy and multiple hippocampal transection by subtemporal approach with disconnection of subiculum and CA1 are considered to improve her seizure. Operative pictures show amygdalotomy and hippocampal transection by subtemporal approach with disconnection of subiculum and CA1 (Figure 13).

Figure 13.

Operative pictures showing amygdalotomy and hippocampal transection by subtemporal approach with disconnection of subiculum and CA1. Upper left: collateral sulcus (aspirator and forceps) and temporal horn opened showing hippocampus and amygdala. Upper right: After amygdalotomy (removal of basolateral nucleus), internal carotid artery and anterior choroidal artery were seen beyond arachnoid membrane. Lower left: After multiple hippocampal transection (blue line), continuous disconnection (yellow line) between CA1 and subiculum from alveus to hippocampal sulcus (lower right schema, red arrow) was done.

Postoperative neuropsychological examinations have improved already 1 day after the operation (Frontal Assessment Battery (FAB) and Hasegawa Dementia Scale (HDS)-Revised in Table 5), and seizure stopped. Figure 14 demonstrated preoperative coronal T2W image and postoperative coronal T2W images, showing amygdalotomy (lower left) and MHT and disconnection between CA1 and subiculum. Two months after surgery, neuropsychological examinations showed slight improvement comparing to preoperative levels (Table 5 except FAB, HDS-R). Seizure control is satisfactory, and only two auras were seen during 1 year after the operation.

Op. date: 2022/1/17.
2018/4/27 (Preop.); WMS-R Verbal Memory. 75 Visual Memory 87 General Memory 74 Attention/concentration 87 Delayed Recall 66 FAB 16/18 TMT A 1′24″ TMT B 2′ MMSE 26/30	2022/3/30 (Postop. 2 months); WMS-R Verbal Memory 78 Visual Memory 89 General Memory 78 Attention/concentration 81 Delayed Recall 83 FAB 18/18 (Preop. 14/18) ADAS-J Cog 5.0/70 TMT A 35′ TMT B 59” HDS-R 30 (Preop 28)

Table 5.

Results of preoperative and postoperative neuropsychological examinations in this patient.

Figure 14.

Preoperative coronal T2W images showing amygdala (upper left) and HS (upper right), and postoperative coronal T2W images showing amygdalotomy (lower left) and MHT and disconnection between CA1 and subiculum.

For disconnection between CA1 and subiculum for HS type 1, it is relatively easy to disconnect, because in HS type 1, CA1 is extremely atrophic and longitudinal limit of CA1 and subiculum is easy to identify.

12. Conclusion

1. Temporal lobe epilepsy, one of the most common forms of epilepsy involving limbic system and intractable for medical treatment, is the most challenging target for neurosurgeons, and various technical improvements are reported. Among these procedures, subtemporal selective amygdalohippocampctomy is seemingly least invasive in terms of neuropsychological function.

2. Technique and the operative and neuropsychological results of subtemporal selective amygdalohippocampectomy for medically intractable temporal lobe epilepsy have been demonstrated.

3. Using anatomical connectivity mapping technique, the subtemporal resection is the least disruptive to long-range connectivity which may explain its better cognitive outcome. These results have a direct impact on understanding the amount and/or type of cognitive deficit postsurgery, which may not be obtainable using local measures of white matter integrity.

4. New operative methods are demonstrated in this text, that is, subtemporal multiple hippocampal transection with or without disconnection between subiculum and CA1. These techniques might be useful to obtain good neuropsychological functions and good seizure control for language-dominant temporal lobe epilepsy.

5. Further operative experiences should be obtained to clarify the usefulness of subtemporal multiple hippocampal transections and disconnection of subiculum and CA1, especially for HS type 1, although randomized clinical trial is difficult to perform.

Abbreviations

TLE

temporal lobe epilepsy

TSSAH

transsylvian selective amygdalohippocampectomy

sSAH

subtemporal amygdalohippocampectomy

PTLE

paradoxical temporal lobe epilepsy

MHT

multiple hippocampal transection

TSMHT

transsylvianl multiple hippocampal transection

STMHT

subtemporal multiple transection