NCOG-13. TOWARDS LONG-TERM MEMORY MAPPING DURING AWAKE CRANIOTOMY: VALIDATION OF A SHORT-FORMAT TASK IN PATIENTS WITH BRAIN TUMORS

Abstract

BACKGROUND: Patients with brain tumors often suffer from memory impairment, a symptom that results in a significant decrement to quality of life. However, unlike sensorimotor and language functions, memory is not mapped during awake craniotomies. Following on research showing that long-term memory processes are operative and can be measured over compact timescales, our goal is to develop a brief probe of memory function that will be suitable for the confines of the OR environment. Here, we report a validation study of a candidate task. METHODS: We designed a novel memory test in which patients studied collections of five objects arranged in a circle. After a 4s unfilled delay, the locations of two objects were swapped and patients attempted to identify these swaps. To vary difficulty, objects included either abstract squiggles or identifiable shapes. Fifty-five patients with supratentorial parenchymal tumors completed a series of standardized neuropsychological tests (including classical memory measures), as well our novel task. Neuropsychological battery performance was compared to performance on the novel paradigm using across-subject correlation analysis. RESULTS: 46/55 patients completed all components of the study. Performance on the squiggle version of the task was strongly correlated with total recall on the standardized battery (r=0.586, p<0.0001), whereas the object version correlated with a variety of domains, including total recall (r=0.4267, p=0.0031), visuospatial (r=0.560, p<0.0001), language (r=0.485, p=0.0006), and attention (r=0.477, p=0.0008). An analysis using moving averages across sequential trials indicated that only four squiggle trials were necessary to obtain a stable estimate of an individual's overall performance, and as low as a single trial for the objects. CONCLUSION: Our novel test provides a practical way to map both long-term memory (squiggles) and global cognition (objects) with low trial numbers, making intraoperative testing during awake craniotomy feasible, and may allow for more reliable preservation of cognition postoperatively.