Frontal Underactivation During Working Memory Processing in Adults With Acute Partial Sleep Deprivation: A Near-Infrared Spectroscopy Study

Figures

• 

  Figure 1. Experimental design of the 0- and 3-back conditions of the n-back paradigm. Each task block began with a 5-s instruction cue, followed by 28 trials (7 target and 21 non-target trials) that were presented in a pseudo-randomized order. In each trial, a digit was displayed for 500 ms, followed by an inter-stimulus interval of 1,000 ms. Participants were required to press the left button of a mouse when the target stimulus appeared but to press the right button for the other stimuli. The target stimulus for the 0-back condition was the digit “0”, and that for the 3-back condition was the digit that was identical to the one presented three trials before.

• 

  Figure 2. Arrangement of the near-infrared spectroscopy (NIRS) channels. The distance between each pair of emitter and detector probes is 3 cm. In accordance with the international 10/20 system, the center of the probe matrix was placed on Fpz, and the probes at the bottom left and right corners were placed around F7 and F8, respectively. Based on previous studies (Okamoto et al., 2004; Koessler et al., 2009), it was estimated that the outermost channels (e.g., ch1-4,13-16) were located around the lateral prefrontal cortices, and the medial channels (e.g., ch7-10) were located around the medial frontopolar cortex and dorsomedial prefrontal cortex.

• 

  Figure 3. Time courses of the working-memory-related (WM-related, i.e., 3-back minus 0-back) [oxy-Hb] and [deoxy-Hb] changes in the Sufficient Sleep group (SS; n = 20). Each n-back block started at 0 s, began with an instruction cue that lasted for 5 s, and ended at 47 s. Channels that exhibited significant WM-related [oxy-Hb] increases (i.e., mean [oxy-Hb] values over the 42-s active task period, excluding the cue period) and those that exhibited significant WM-related [deoxy-Hb] decreases (i.e., mean [deoxy-Hb] values over the 42-s active task period, excluding the cue period) are denoted by red (ch1-4,13-16; ps < 0.05; FDR-corrected) and blue square borders (ch2-16; ps < 0.05; FDR-corrected), respectively. Lines and shading indicate the mean ± 1 standard error.

• 

  Figure 4. Time courses of the working-memory-related (WM-related, i.e., 3-back minus 0-back) [oxy-Hb] and [deoxy-Hb] changes in the Insufficient Sleep group (IS; n = 20). Each n-back block started at 0 s, began with an instruction cue that lasted for 5 s, and ended at 47 s. None of the channels exhibited significant WM-related [oxy-Hb] increases (i.e., mean [oxy-Hb] values over the 42-s active task period, excluding the cue period) or significant WM-related [deoxy-Hb] decreases (i.e., mean [deoxy-Hb] values over the 42-s active task period, excluding the cue period) (ps < 0.05; FDR-corrected). Lines and shading indicate the mean ± 1 standard error.

• 

  Figure 5. t-Contrast maps of the mean working-memory-related (WM-related, i.e., 3-back > 0-back) [oxy-Hb] and [deoxy-Hb] changes in the (A) Sufficient Sleep (SS; n = 20) and (B) Insufficient Sleep (IS; n = 20) groups and (C) between the SS and IS groups. For the top panels of (A,B), red and blue indicate WM-related [oxy-Hb] increases and decreases, respectively; for the bottom panels of (A,B), red and blue indicate WM-related [deoxy-Hb] increases and decreases, respectively. In addition, for the top panel of (C), red indicates greater WM-related [oxy-Hb] increases in the SS group; for the bottom panel of (C), blue indicates greater WM-related [deoxy-Hb] decreases in the SS group.

• 

  Figure 6. Correlations between sleep duration the night before and mean working-memory-related (WM-related, i.e., 3-back > 0-back) [oxy-Hb] changes on the experimental day. Pearson's correlation coefficient (r) was calculated for each channel, and a linear regression line was fitted to each channel. Channels that exhibited significant correlations are denoted by square borders (ch1-16; FDR-corrected; ps < 0.05; two-tailed).

• 

  Figure 7. Correlations between sleep duration the night before and mean working-memory-related (WM-related, i.e., 3-back > 0-back) [deoxy-Hb] changes on the experimental day. Pearson's correlation coefficient (r) was calculated for each channel, and a linear regression line was fitted to each channel. Channels that exhibited marginally significant correlations are denoted by square borders (ch10-11,14-16; uncorrected; ps < 0.05; two-tailed).