Elsevier

Clinical Neurophysiology

Volume 112, Issue 9, September 2001, Pages 1672-1675
Clinical Neurophysiology

Transcranial magnetic stimulation of left prefrontal cortex impairs working memory

https://doi.org/10.1016/S1388-2457(01)00606-XGet rights and content

Abstract

Objectives: Several lines of evidence suggest that the prefrontal cortex is involved in working memory. Our goal was to determine whether transient functional disruption of the dorsolateral prefrontal cortex (DLPFC) would impair performance in a sequential-letter working memory task.

Methods: Subjects were shown sequences of letters and asked to state whether the letter just displayed was the same as the one presented 3-back. Single-pulse transcranial magnetic stimulation (TMS) was applied over the DLPFC between letter presentations.

Results: TMS applied over the left DLPFC resulted in increased errors relative to no TMS controls. TMS over the right DLPFC did not alter working memory performance.

Conclusion: Our results indicate that the left prefrontal cortex has a crucial role in at least one type of working memory.

Introduction

Working memory refers to temporary storage and manipulation of the information necessary for complex tasks such as language comprehension, learning, and reasoning (Baddeley, 1992). Fuster et al. (1982) found that some neurons of the prefrontal cortex increase their firing when a cue is presented and continue to fire during a delay period after the cue disappears. Functional magnetic resonance imaging (fMRI) and positron emission tomography (PET) studies indicate that the frontal cortex plays a crucial role during working memory tasks (e.g. Roland, 1984, Paulesu et al., 1993, Jonides et al., 1993, Petrides et al., 1993a, Petrides et al., 1993b, Cohen et al., 1994). PET and fMRI studies have shown increased metabolic activity in the frontal lobes during working memory tasks, with other cortical areas also activated depending on the task involved (e.g. Berman et al., 1995, Smith et al., 1996, Salmon et al., 1996, D'Esposito et al., 1995). The activation of frontal cortex appears to be proportional to working memory demands and not ‘mental effort’ more generally (Barch et al., 1997). Using EEG techniques with high spatial resolution, Gevins et al. (1996) identified several frontally localized waveforms modulated by working memory task manipulations. A left-lateralized slow frontal positivity with a mean peak latency of 450 ms (P450) was larger in both spatial and verbal memory tasks than in the respective controls.

Single-pulse transcranial magnetic stimulation (TMS) can transiently disrupt the function of restricted regions of cortex. For example, TMS over sensory cortex can decrease perception of cutaneous stimuli delivered to the fingers of the contralateral hand (Cohen et al., 1991, Seyal et al., 1992) for up to 500 ms after the TMS pulse (Seyal et al., 1997). Repetitive TMS (rTMS) of the frontal cortex has been shown to increase errors in a visuospatial delayed-recall task when stimulation was applied throughout the entire delay period, but not with a shorter duration of stimulus (Pascual-Leone and Hallett, 1994). TMS of human cortex causes brief disruption of cortical activity and can therefore provide information on dynamic cortical processes with sub-second temporal resolution. Single-pulse TMS can be safely used in normal human subjects without the risks inherent with rTMS (Wassermann, 1998).

We proposed that under certain conditions single-pulse TMS should be effective in disrupting verbal working memory. First, we targeted the pulse to an approximate time and location where Gevins et al. (1996) found EEG evidence of dorsolateral prefrontal cortex (DLPFC) activity in a verbal working memory task. Second, we ensured that subjects were engaged in a task with a high working memory load by having them participate in a relatively difficult ‘3-back’ working memory task. This ‘3-back’ sequential letter matching task activates the DLPFC (Cohen et al., 1997) Third, we explored a number of regions on the left frontal scalp of each subject to find the location where TMS appeared to have the greatest effect on task performance.

Section snippets

Methods

Nine healthy human subjects were tested. The age range was 23–59 years (mean 34 years). Eight were strongly right-handed; one was strongly left-handed. Subjects gave informed consent, and the local Human Subjects Review Committee approved the study.

Subjects were presented with a pseudo-random set of 33 letters (A–J). Letters were displayed serially on a backlit LCD screen for 30 ms every 2 s. Subjects were required to state if the letter just presented was the same as the letter presented

Results

Seven subjects completed the entire experiment; two completed 6 of 8 sets. In all, there were 34 sets of data following TMS (1020 responses) and the same number of no TMS controls for each hemisphere. Fig. 1 shows the errors made by each subject during the TMS condition and the corresponding control condition.

With left frontal scalp stimulation, significantly more errors were made following TMS than in the corresponding controls (P=0.008). There were 183 (17.9%) incorrect responses in the left

Discussion

In this study, there was a significant increase in task errors related to TMS applied over the left DLPFC relative to the control condition. This degradation in task performance is likely related to transient functional inactivation of the left DLPFC by TMS. Right prefrontal cortex stimulation resulted in no significant change in working memory performance relative to the control condition.

The effect of single pulse TMS on working memory in this study is less pronounced than that reported with

Conclusion

The results of this study indicate that working memory performance, in a sequential-letter-matching task, is impaired by TMS-induced functional inactivation of left prefrontal cortex. This effect occurs during the period when EEG evidence suggests that this region of cortex is engaged in a working memory task. PET and fMRI provide localizing information but have relatively poor temporal resolution as these techniques are dependent on hemodynamic changes that are delayed and temporally dispersed

Acknowledgements

This project was supported by the Predoctoral Research and Enrichment Fellowship, University of California, Davis, School of Medicine (Dr Mull).

References (27)

Cited by (119)

  • Transcranial magnetic stimulation: Neurophysiological and clinical applications

    2019, Handbook of Clinical Neurology
    Citation Excerpt :

    The authors found evidence of a subtle double dissociation: putatively reducing left DLPFC activity tended to both decrease verbal performance while increasing spatial accuracy, while the opposite was observed when the right DLPFC was stimulated (Fried et al., 2014). Some TMS studies have found similar evidence of a verbal/spatial hemispheric specialization (Mull and Seyal, 2001; Sandrini et al., 2008), while others have failed to find such a distinction (Oliveri et al., 2001; Mottaghy et al., 2002a, 2003a; Rami et al., 2003; Postle et al., 2006). To reconcile these inconsistencies, it has been suggested that the PFC is organized by both domain and the type of process (Sandrini et al., 2008; Zanto et al., 2011).

  • Effects of TDCS dosage on working memory in healthy participants

    2018, Brain Stimulation
    Citation Excerpt :

    The total amount of charge (millicoulombs) delivered during the experiment session for each stimulation condition was: Off-0mC; Sham1-134 mC; Sham2-66 mC; 1mA-930mC; 2mA-1860mC. The visual 3-back working memory task, adapted from Mull et al. [43], was administered using Inquisit 4 software (Version 4, Millisecond Software). This task required participants to press a key (spacebar) when a displayed letter matched a letter shown three trials previously.

  • Interactive Benefits from Switching Electrical to Magnetic Muscle Stimulation

    2023, UIST 2023 - Proceedings of the 36th Annual ACM Symposium on User Interface Software and Technology
View all citing articles on Scopus
View full text