November 26, 2012

Sleep Deprivation and the Stroop Task

I've decided to start this post off with a small test. See how fast you can identify the font colour of the words below ... and ... GO!

So - how did you do? Did you find it difficult to ignore what the word said and focus instead on the actual colour of the font? If you did, don't worry - it is completely normal to find it difficult to inhibit your automatic response to read the word. What you just did is a common task known as the Stroop task, devised by John Ridley Stroop in 1935. The basis of this task involves the use of incongruent stimuli (i.e., using the name of a colour that is printed in a colour not denoted by the name) to cause what is fondly known as the Stroop effect - where naming the font colour takes longer (measured by reaction time, or RT) and is more prone to errors than when the font colour matches the written word itself. Since Stroop's initial publication on the Stroop effect (see: Studies of interference in serial verbal reactions by J.R. Stroop, 1935), the Stroop task has been widely replicated and is currently one of the most commonly known and used tasks in research and clinical practice worldwide. And so from this brief background on the Stroop task, I now segue into a recent paper published in Brain and Cognition in 2011 - a paper that not only involves said Stroop task, but also incorporates an (indirect) aspect of my Masters thesis: sleep deprivation.

Cain, S.W., Silva, E.J., Chang, A.M., Ronda, J.M. and Duffy, J.F. (2011) One night of sleep deprivation affects reaction time, but not interference or facilitation in a Stroop task. Brain and Cognition 76: 37-42.

Goals of the study?
In this study, Cain and colleagues aimed to study the effects of prolonged wakefulness on Stroop task performance by placing thirty healthy participants in a 40-hour "constant routine" (CR) protocol, during which the participants remained awake in constant conditions and performed a Stroop task every two hours. Several studies prior to this have examined the performance of the Stroop task under acute sleep deprivation conditions, but the results of these studies were not able to differentiate between the general effects of sleep deprivation on performance and the effects specifically related to Stroop interference. Moreover, these previous studies were unable to determine whether the sleep loss-related decrements in Stroop task performance were due to slower RTs or whether executive function (and the prefrontal cortex) were also affected by sleep deprivation.

Cain and his research team hypothesized that sleep deprivation would affect RT and error rate but not interference in the Stroop task. This would indicate that attention and vigilance may be more sensitive to sleep loss than the executive functions involved in Stroop interference.

- Three 24-hour baseline days (16 hours of wakefulness and 8 hours of bed rest in the dark)
- 40-hour CR (see above)
- Non-24-hour sleep/wake cycle (designed to determine the period of their circadian timing system)
Interference = mean RTs of incongruent trials - mean RTs of neutral trials (reflects additional time needed to respond when reading the word interferes with responding to the colour)
Facilitation = mean RTs of congruent trials - mean RTs of neutral trials (reflects the increased speed of reaction when colour and word matched)

There were several significant results that came out of this study. I will list them in the order that they were presented in the paper.

1. There was a significant main effect of time awake on RT, such that RT slowed with increasing time awake.
2. There was a significant main effect of trial type on RT, such that incongruent trials showed the slowest RT, congruent trials showed the fastest RT and neutral trials showed intermediate RTs.
3. There was a significant main effect of time awake on the error rate of responses, such that there were more errors with increasing time awake.
4. There was a significant main effect of trial type on error rate, such that the incongruent trials had the most errors made, the congruent trials had the fewest errors made, and the neutral trials had an intermediate number of errors made.
5. For RT, significant differences between incongruent and neutral trials were observed at 6, 8, 12, 14, 16, 22, 34, 36 and 38 hours awake.

There were also several non-significant results that came out of this study. Again, I will list them in the order that they were presented in the paper.

1. There was no significant interaction between the main effects of time awake and trial type on RT.
2. There was no significant interaction between the main effects of time awake and trial type on error rates of responses.
3. There was no significant main effect of time awake on indexes of interference or facilitation.

What do these results mean?
An interesting pattern of performance was found when analyzing the results (and this is what REALLY interested me!).
Performance on all three trial types was stable across the first 14-16 hours of CR, which corresponded to the usual waking hours of the participants. For the next 10-12 hours following this period (which corresponded to the usual sleeping hours of the participants), both RTs and error rates showed increases, suggesting that there was no apparent speed/accuracy trade-off. However, RTs and error rates on all three trial types stabilized after about 24 hours awake, and (fascinatingly) rebounded slightly between hours 28 and 36! This pattern of performance rebound in the morning after a night awake may reflect an underlying circadian rhythm in sleep-wake propensity, as studies using other types of cognitive performance have demonstrated a similar pattern.
Both indexes of interference and facilitation used were not significantly affected by sleep deprivation. Therefore, the executive function of inhibiting the prepotent response of responding to the word did not appear to be very sensitive to one night of sleep loss. It may be, however, that one night of sleep loss was merely insufficient to affect these processes. Furthermore, these findings may have arisen from a problem known as "task impurity", in which a particular task may involve the interaction of executive and non-executive components, both of which influence performance. It is thus posited that sleep loss may affect overall performance on the task via either one or more of these components.

As expected, RTs were slowest for incongruent trials, fastest for congruent trials and intermediate for neutral trials. Also expected were the errors made during incongruent (most errors), congruent (least errors) and neutral trials (intermediate errors). It was thus concluded that under constant environmental and behavioural conditions, one night of sleep loss slows down RT and increases error rates in Stroop performance, but appears to have no significant impact on Stroop interference or facilitation in healthy, young individuals.

So now that my scientific rambles are all out there for you to read, the most important question is: what is the importance of all of this? I mean, it's not every day that you'd be kept awake for 40 hours in a room with constant conditions and told to perform the Stroop task every two hours, right?

... And to that I say, of course this study is important! For one, it gives me some background as to what to expect during my data collection and analyses. Mind you, my participants won't be forced to stay awake, but sleep deprivation will most definitely be a side effect experienced by the experimental group. Moreover, although no interference effects were found in this study, my study may show otherwise as my participants may suffer from days of disrupted sleep. I am therefore very much looking forward to comparing and contrasting my findings to those presented in this paper. Lastly and most importantly (for you readers out there, wherever you may be), this study gives you an idea of the effects acute sleep deprivation has on your mental and physiological processes. Though you may not consciously feel impaired, this study shows that your inhibition processes, for one, are definitely taking a hit. So the next time you're debating whether or not to pull an all-nighter to finish your 50% essay, think otherwise!

And with that final piece of advice, I'm signing off. I'm headed to the scanner to help scan Ken Swift tonight, a world-famous, highly influential B-boy and an honourary guest at York University's own 3T scanner. Updates on my meet and greet with Ken Swift a.k.a. "The Epitome of a B-Boy" to come!