Cognitive Processes in the Breakfast Task: Planning and ...

1 Cognitive Processes in the Breakfast Task: Planning and MonitoringNathan S. RoseRotman Research Institute at Baycrest, Toronto, CanadaLin Luo and Ellen BialystokYork UniversityAlexandra Hering, Karen Lau, and Fergus I. M. CraikRotman Research Institute at Baycrest, Toronto, CanadaThe Breakfast Task (Craik & Bialystok, 2006) is a computerized task that simulates the Planning andmonitoring requirements involved in cooking Breakfast , an everyday activity important for functionalindependence. In Experiment 1, 28 adults performed the Breakfast Task, and outcome measures wereexamined with principal component analysis to elucidate the structure of Cognitive Processes underlyingperformance. Analyses revealed a 2-component structure which putatively captured global Planning andlocal monitoring abilities. In Experiment 2, the structure of Breakfast Task performance was cross-validated on a new sample of 59 healthy older adults who also performed tests assessing workingmemory, processing speed, inhibition, reasoning and prospective memory.

2 Factor analyses showed thatthe global Planning component from the Breakfast Task was significantly correlated with individualdifferences in executive functions but the local monitoring component was independent of such func-tions. The Breakfast Task provides a fast, enjoyable, and lifelike assessment of complex everydayplanning and monitoring, and their underlying Processes such as working memory and : Planning , monitoring, working memory, prospective memory, executive processesThe Cognitive Processes underlying many everyday activities,such as running errands, shopping for groceries, or preparing ameal, are surprisingly complex. To complete such activities it isnecessary to formulate a sequence of actions to achieve the goals,store and update the action plan in working memory, and con-sciously monitor and coordinate the execution of subtasks. Theseaspects of Cognitive Processes are often grouped under the um-brella termplanning(Morris & Ward, 2005).

3 Efficient planningclearly depends on a multitude of Cognitive Processes , but therelative contribution of specific Processes is less clear, and theirapparent involvement may depend both on the particular task andon how Planning is studies investigating Planning behaviour typicallyadopt one of two general approaches (Ward & Morris, 2005). Thefirst approach is represented by tower tasks, such as Tower ofHanoi and its variants ( ,Owen, 2005), in which participants arerequired to solve an unfamiliar problem following a set of restric-tions. This approach has the advantage of experimental control, butis often criticised for its limited applicability to everyday situa-tions. The second approach, in contrast, uses tasks in which thegoals and contexts are common in everyday life. Examples of thismethod include errand tasks and their variants (Burgess, Simons,Coates, & Channon, 2005), in which participants are given a seriesof tasks to perform in a set order, typically in real-life approach has the advantage of better ecological validity, butthe complexity of such tasks reduces the ability to specify thecognitive components.

4 The present study aims to give a fullerdescription of the Cognitive Processes involved in Planning byusing a simulated real-life task that has both ecological validityand experimental task in question is the Breakfast Task (Craik & Bialystok,2006), which is a computerized task that simulates the planningand task management requirements in a common everyday situa-tion: cooking Breakfast . The task is described in detail later, but inoverview the participant s task was to cook five Breakfast foodsin such a way that the five foods were not over- or undercooked,and were ready to serve at the same time. The task was imple-mented on a touch-screen computer, and participants started andstopped cooking each food by pressing the START and STOP buttons next to the icon indicating that food (seeFigure 1).Cooking progress of each food was indicated by a countdowntimer running in real time. Between starting and stopping cooking,the participant also carried out a table-setting task as a backgroundThis article was published Online First May 4, S.

5 Rose, Rotman Research Institute at Baycrest, Toronto, Can-ada; Lin Luo and Ellen Bialystok, Department of Psychology, York Uni-versity; Alexandra Hering, Karen Lau, and Fergus I. M. Craik, RotmanResearch Institute at of the authors has any financial interest or will receive anyfinancial benefit from the work reported in this article. This work wassupported by Grant 8261-03 to Fergus I. M. Craik from the NaturalSciences and Engineering Research Council of Canada, Grant MOP57482to Ellen Bialystok and Fergus I. M. Craik from the Canadian Institutes ofHealth Research, and a Kickstart Grant from the Centre for Brain Fitnessto Nathan S. Rose and Fergus I. M. Craik. We thank Sharyn Kreuger, JaneLogan, Ashley Morgan, Evelyn Zastepa, Karen Lau, Yunjo Lee, JennaWare, and Brittany Murphy for their help with data collection and concerning this article should be addressed to FergusI. M. Craik,Rotman Research Institute at Baycrest, 3560 Bathurst Street,Toronto, Ontario M6A 2E1, Canada.

6 Journal of Experimental Psychology / Revue canadienne de psychologie exp rimentale 2015 Canadian Psychological Association2015, Vol. 69, No. 3, 252 2631196-1961/15/$ The overall task was to coordinate and manage cookingthe foods, while setting as many places as Breakfast Task clearly demands good multitasking abilitiesto perform it efficiently, and previous studies have provided mod-els of Planning in a multitasking context. As one example,Burgess, Veitch, de Lacy Costello, and Shallice (2000)studied theperformance of 60 brain-damaged patients and 60 controls on theGreenwich Test, which consists of three complex open-ended tasksto be performed in 10 min. Their data suggested a three-factormodel consisting of memory, Planning , and intention, with plan-ning and intention (including prospective memory) being depen-dent on good learning and memory for the rules governing the subsequent study byLogie, Trawley, and Law (2011)followedup theBurgess et al.

7 (2000)findings but used healthy young adultsas participants. They used a multiple-errands task performed in avirtual environment: a 3-D model of a four-story colleagues (2011)confirmed the three-factor structure ofmemory, Planning , and intention proposed by Burgess and col-leagues, although memory was not as crucial as a precursor forplanning as in the earlier study. Additionally, Logie and colleaguesfound that a measure of spatial working memory was a strongpredictor of success in their task. Interestingly they used theBreakfast Task as their measure of prospective memory but foundthat Breakfast Task performance did not predict success on theirmultiple-errand a tentative summary statement, it may be suggested that thefactor structure of different multitasking situations will vary de-pending on the precise needs of each particular task. Thus, forexample, the Greenwich Test used byBurgess et al. (2000)re-quires substantial learning of complex rules, so Planning andintention depended on adequate learning and memory.

8 The taskused byLogie et al. (2011)required retention of visuospatialinformation so it makes sense that visual working memory was astrong predictor of success in their experiment. In the case of theBreakfast Task we predicted that good performance would dependon the ability to formulate and execute a plan, and to maintain theoverall goal in mind while monitoring and executing the and Bialystok (2006)proposed that the Cognitive abilitiesinvolved in achieving these goals include working memory, tohold both the general plan and the progress of each food in mind,task switching to switch efficiently between cooking and tablesetting, and prospective memory, to remember to start and stopfoods Breakfast Task measures are sensitive to factors known toaffect executive functioning and prospective memory, such ashealthy aging (Craik & Bialystok, 2006; Kerr, 1991), bilingual-ism (Craik & Bialystok, 2006), and neurological pathologies(Bialystok,Craik, & Stefurak, 2008;West, McNerney, & Krauss,2007), providing some support for their validity.

9 However, theinterpretation of the measures was substantially intuitive. No directevidence is yet available to establish the relationship between eachmeasure and its proposed underlying Cognitive Processes . Theprimary purpose of the present study was thus to provide a moreprincipled description of the Cognitive Processes involved in theBreakfast Task itself; a secondary purpose was the more generalFigure sample screen for the one-screen condition of the Breakfast Task, showing foods in the processof being cooked. Neither the toast nor the coffee has been started, so their cooking indicators are not yet eggs and pancakes have each been cooking for a few seconds, and the sausages have been cooking for over1 min (Reproduced with permission fromCraik & Bialystok, 2006). See the online article for the color versionof this Processes IN THE Breakfast TASKone of providing more information on the Cognitive Processes thatcontribute to efficient Planning and task management the first experiment we tested 28 adults aged 39 68 years onthe Breakfast Task.

10 We performed principal component analysis(PCA) on the set of Breakfast measures to examine the structure ofperformance. In the second experiment we tested an independentsample of 59 healthy older adults on the task to cross-validate thestructure of Breakfast Task performance and on a range of testsof Cognitive functions. These tests focused on the measurement ofexecutive functions (EF) and prospective memory (PM), the maincandidate Processes underlying complex Planning activities. Justas Planning is not a unitary construct, neither should executivefunctions and prospective memory be regarded as unitary con-structs. We therefore included multiple tests and measures toreflect different aspects of these abilities. We performed explor-atory factor analyses on the sets of EF and PM measures to extractprocessing components tapped by these measures, and then exam-ined the relations between these processing components and thebreakfast measures with correlational analyses.