Functional Imaging of Working Memory

This report summarizes a study investigating the effects of 24 hours of total sleep deprivation on working memory performance and associated brain activation patterns in young, healthy adults. The study employed functional magnetic resonance imaging (fMRI) to measure brain activity during the performance of two working memory tasks: a low-load task (LTR) and a high-load task (PLUS). The results indicated that sleep deprivation led to slower reaction times and increased subjective sleepiness. Furthermore, sleep deprivation resulted in reduced deactivation in brain regions associated with the default mode network (anterior medial frontal and posterior cingulate cortices), suggesting impaired cognitive efficiency. Increased activation was observed in the left dorsolateral prefrontal cortex, potentially reflecting compensatory effort. The study concludes that efficient working memory processing is associated with up-regulation of task-related brain regions, and that sleep deprivation impairs this efficiency.

Notes for paper: Functional Imaging of Working Memory after 24 Hr of Total Sleep
Deprivation (Chee & Choo, 2004).
Aim(s)
 To examine the interaction between task complexities and state that modulates
cortical activation in young, healthy adults during working memory tasks.
Participant
 n= 14 right handed, young, healthy adults (mean age- 23 years).
 male- 9, female- 5.
Study Design
 Volunteers with good sleep patterns.
 Average sleep of 7.2 +/- 0.9 hours per night.
 No volunteers included with history of insomnia, daytime sleepiness, psychiatric
illness, narcolepsy or addicts.
Protocol
 SD and RW scanning of subjects conducted within 1 week.
 Subjects abstained from caffeine, smoking, for 24 hr before scanning. Alcohol was
disallowed.
 Engaged in non-strenuous activities like conversing and video watching.
 Epworth sleepiness scale (ESS) used to rate sleepiness and simple reaction time task
(SRT) performed.
 180 trials executed.
Measures

 3T Allegra magnetic resonance imaging
 T1-weighted three-dimensional-MPRAGE imaging
 Motion correction using PACE (Siemens)
 Mean intensity normalization
 LTRRW, LTRSD, PLUSRW, PLUSSD used for functional analysis
 Region of interest based analysis
Results
 Behavioral results-
 Increased ESS showed greater subjective sense of sleepiness after SD.
 Responses omitted after SD not for PLUS but for LTR.
 Slow RT for both PLUS and LTR.
 RW activation-
 Left hemisphere dominant activation observed in LTR in prefrontal (pre-
central regions, insula, Brodman’s area and thalamus).
 Areas showed more activation in PLUS than LTR.
 These areas were left prefrontal region around left insula, left inferior parietal
lobule, middle frontal gyrus and bilateral thalamus.
 No region more activated in LTR compared to PLUS.
 Cortical deactivation lesser for LTR than PLUS in RW.
 Activation after SD when compared to RW-
 PLUS and LTR showed activation in overlapping regions after SD than during
RW.
 LTR and PLUS elicited smaller task related BOLD signal by Voxel-by-voxel
analysis in the parietal region when compared to RW.

 Larger post SD increase in BOLD signal observed in left dorsolateral
prefrontal cortex near middle frontal gyrus compared to RW.
 Prefrontal activation by LTR showed no significant difference after SD.
 Reduced deactivation after SD observed in left posterior cingulate and anterior
medial frontal regions in voxel-by-voxel contrasts.
 ROI-based analyses showed effects of state and in posterior cingulated and
anterior medial frontal regions. No significant state-by-task interaction found.
 RT inversely correlated with the anterior medial frontal region deactivation
 RT did not correlate with change in BOLD signal in posterior cingulated, left
dorsolateral prefrontal, and thalamus or parietal regions.
Interpretation
 Greater parietal and prefrontal activation shown by PLUS than LTR in both states that
matched the notion that when load of working memory is not excessive, additional
processing resources are engaged by manipulation.
 Left dorsomedial thalamus activation was more associated with PLUS. Previous
experiments demonstrated activation of thalamus with increased working memory
demand.
 The thalamic activation increase with prefrontal activation signified the existence of
reciprocal connections between prefrontal cortex and dorsomedial thalamic nucleus.
Sustained attention also increased thalamic activation. These mechanisms contributed
to task related differences in thalamic activation.
 Deactivation of BOLD signal relative to baseline during task performance was
pronounced with complex task in both the states. Deactivated anterior medial frontal
and posterior cingulate regions were part of activated default network during passive
conditions.

 Inverse relationship between RT and deactivation magnitude suggested greater
deactivation is related to efficient task performance. After SD recruitment of cognitive
resources are required in diminished capacity to engage in goal directed behaviours.
 The results of functional imaging prove that efficient processing of the brain is
associated with an up-regulation in the activation of task related regions.
 Older respondents who show decline in activation of frontal regions of the brain. They
exhibit poor retrieval of memory. They displayed bilateral or non-selective increase in
activation when environmental support that facilitates encoding was provided.
Mnemonic performances improved when compared to younger counterparts.
 Reduction in activation of occipital lobe is associated with impairment in sensory
processing in elderly patients.