Characterization of a Factual Knowledge–Associated Brain Memory Area by Functional Magnetic Resonance Imaging and Implementation in Tumor and Epilepsy Brain Surgery

Background

The role of the dominant hippocampus in verbal and episodic memory recall has been extensively investigated. However, there are categories of memory estimated independent of hippocampal function. In particular, factual knowledge is solely constituted from written or spoken input independent on undertaken experience, but the temporal structures involved are not well defined.

Methods

In this retrospective study, we mapped the activity of factual knowledge recall using functional magnetic resonance imaging (fMRI) in 60 patients scheduled for tumor or epilepsy surgery and in 18 patients resected within or close to the left mesial temporal lobe. Memory tests were performed preoperatively and postoperatively to investigate deficit symptoms in factual knowledge retrieval.

Results

We found a highly reproducible localized brain area within the parahippocampal gyrus, in the depth of the dorsal collateral sulcus, which was constantly involved in factual knowledge retrieval during fMRI tasks in 93% of patients (56/60) during recall of capital cities or multiplication tasks, in two thirds of the patients mainly on the dominant side. In addition, all 18 patients after surgery within this area showed postoperative factual knowledge deficits (56% permanent, 44% transitory), memory prognosis significantly correlating with the first day's factual knowledge retrieval task results.

Conclusions

Using fMRI, we visualized a circumscribed brain area within the parahippocampal gyrus during factual knowledge retrieval, the lesioning of which led to factual knowledge deficits. Thus, impairment of factual knowledge retrieval may occur if resection is extended into the parahippocampal gyrus. Preoperative fMRI memory maps may contribute to avoiding such deficits.

Introduction

Various types of amnesia are observed in patients with temporomesial brain tumors, stroke, trauma, or after tumor resection or epilepsy surgery in the dominant temporomesial structures. Two types of amnesia have been classified: retrograde amnesia and anterograde amnesia, comprising loss of previously acquired memory and loss of the ability to build up or keep new memories. In patients with retrograde amnesia, different parts of memory content can be affected. This factor led to a classification of different memory subtypes. Besides episodic or autobiographic memory, which summarizes all experiences obtained in particular periods, factual knowledge, as part of semantic memory, is constituted from facts exclusively derived from written or spoken input (e.g., facts learned in school such as multiplication tables or capitals of countries). The distinction between episodic memory and semantic memory was introduced by Tulving.¹ Sometimes, the denotation of words is also considered as part of memory. The cortical representation of word meaning is confined to areas that have been previously described2, 3, 4 and its irritation was observed after surgery of tumors adjacent to these areas.

Numerous studies have described patients with lesions, who had amnesia caused by stroke or infection. However, in these cases, the extent of the lesion is usually large and its borders are often vague. Rempel-Clower et al.⁵ and Squire et al.⁶ described loss of factual knowledge as shown by tests involving public events and famous faces in 3 patients with brain infarction in the medial temporal lobe. Particularly affected were the dominant hippocampus and entorhinal cortex as determined by postmortem studies. De Renzi et al.⁷ described a patient with destruction of the dominant inferior left temporal lobe, who showed an impaired factual knowledge although autobiographic memory was intact.

Yasuda et al.⁸ reported on a patient who had a poor knowledge of public events (including city names) as a result of radionecrosis of the medial temporal lobes. Grossi et al.⁹ described a case of traumatic brain injury with loss of semantic capabilities such as geographic knowledge and multiplication. Markowitsch et al.¹⁰ also described loss of world knowledge as a result of a traumatic brain injury, although autobiographic memory remained normal. Grewal¹¹ described a left temporolateral infarction exclusively confined to the hippocampus and parahippocampal gyrus, which caused severe loss of factual knowledge. Eslinger et al.¹² reported on a patient with a lesion in the left mesial temporal area with loss of knowledge about public events and names. Reed et al.¹³ described bihemispheric inferior temporal lobe lesions, which caused retrograde amnesia for facts and events. Kapur¹⁴ noted that anterior and inferior temporal lobe structures are particularly important for memory recall.

All the lesions involved in these cases were relatively extensive and their boundaries were not completely clear because of the nature of the lesions. Also, some studies of lesions that caused retrograde amnesia, including worsening of factual knowledge, which were caused by surgery (mostly to cure medically untreatable epilepsy), have been reported.15, 16, 17

In addition, functional magnetic resonance imaging (fMRI) studies18, 19, 20 also indicate involvement of the parahippocampal gyrus and the hippocampus in memory processes, but most of these studies were not confined to verbal memory and specifically factual knowledge. Generally, we share the experience that fMRI imaging of the hippocampus,²⁰ contrary to magnetoencephalography imaging²¹ with verbal material, is successful in only a few patients. One possible explanation for this observation may be the permanent activity of the hippocampus, also in the off intervals used during block design fMRI evaluation. However, in our experience, this is not a problem when visualizing activity in the parahippocampal gyrus.

Therefore, the goal of this study was to assemble evidence for an area that is active and indispensable for recall of factual memory content.