Brief Report: Episodic Foresight in Autism Spectrum Disorder

Abstract

Episodic foresight (EpF) or, the ability to imagine the future and use such imagination to guide our actions, is an important aspect of cognition that has not yet been explored in children with autism spectrum disorder (ASD). This is despite its proposed links with theory of mind (ToM) and executive function (EF), two areas found to be impaired in ASD. Twenty-five children with ASD (M = 5 years, 10 months; 22 male) and 25 mental-age-matched typically developing children (M = 4 years, 10 months; 22 male) completed a series of EpF, ToM, and EF tasks. Significant group differences were detected on several EpF tasks suggesting that children with ASD show impairments in thinking about their future selves.

Appendix

Episodic Foresight Tasks

Sequencing (Modified from Busby Grant and Suddendorf 2009)

Children were shown a timeline, which consisted of a colored board with two Velcro pieces placed at either end of the timeline. Children were told that they were going to play a game involving “how long until things happen in the future.” At one end of the timeline was a small black and white figure depicting a child-size person, and at the other end was a larger black and white figure, depicting an adult (although Busby Grant and Suddendorf’s task included 3 time points, our pilot testing revealed that this was too complex for 3 year-olds; thus, only two time points were retained). The experimenter explained the timeline to the child (“this is a picture of a person as big as you, a 3/4/5-year-old, just like you are right now. We will put things here that are going to happen in a little time, like how big you are right now”). The experimenter also explained to children that they were to place items next to the appropriately-sized person. There were six target events, three for each of the two differently-sized people. For the child-sized person, these included “going home” (house), “having dinner” (plate with food), and “playing at the park” (playground)/“playing in the snow” (snowman; for use during winter testing). For the adult-sized person, these included “get married” (bride and groom), “go to work” (person with briefcase), and “cook dinner” (people cooking). The child was then asked to name each event and place it on the board. If children placed it midway between the two points, they were told that they could only choose one. Each card was removed before presenting the next card. The events were presented in random order, with the requirement that no more than two items from the same category be presented in a row. The number of correct placements was recorded (Total score: range = 0–6). Reliability coding resulted in an interclass correlation coefficient = 1.0.

Picture Book (Atance and Meltzoff 2005)

Children were shown four large color photographs, one at a time, of four “trip” destinations: a waterfall, a desert with a long road, a mountain, and a river with large flat rocks. Children were told to pretend they were going to visit the location in the picture. They were then shown three pictures of items they could choose to bring with them. Children were asked: “Which one of these do you need to bring with you to (e.g., walk close to the waterfall)?”. After selecting one item to bring with them, children were asked “How come you need to bring X?”. The items for each photograph were Band-Aids, fish, pillow (river); water, present, plant (desert); raincoat, rocks, money (waterfall); and lunch, bowl, grass (mountain). Children received a score of 1 for each correct item chosen (i.e., Band-Aids, water, raincoat, and lunch). The correct items were those that could address a future state (e.g., water to satisfy potential thirst). Children also received a score of 1 for each correct explanation (e.g., “because I might get thirsty”). Their explanations were coded as correct only if they contained both a future referent (e.g., gonna, might, will) and a state referent (e.g., hungry, thirsty, hurt). For each trial, scores ranged from 0 to 2 (Total score: range = 0–8). Reliability coding resulted in an interclass correlation coefficient = .97, p < .001, indicating significant agreement.

Grocery/Beach (Hudson et al. 1995)

This task assessed children’s ability to discuss preparatory (planning) activities for two familiar events. After hearing a model plan for going on a picnic that emphasized preparatory activities, children were asked to tell the experimenter plans for the familiar activities of going to the beach and going grocery shopping. Children’s responses were coded for the number of actions generated. A response was considered an action if it contained a verb (e.g., “you swim”, “you buckle your seatbelt”), or if it included an optional or conditional action (e.g., “you could buy strawberries”, “you might go swimming”). Repetitions of previously mentioned actions and nonsensical unrelated information (e.g., “I like dinosaurs”) were not coded as actions. Reliability coding resulted in an interclass correlation coefficient = .98, p < .001, indicating significant agreement.

Tomorrow (Busby and Suddendorf 2005)

This task assessed children’s ability to report things that they “will” and “will not” do tomorrow. Children were introduced to “Scruffles,” a stuffed dog, who asked them “Can you tell me something you are going to do tomorrow?”, and “Can you tell me something you are not going to do tomorrow?”. Each question was asked twice. Children received a score of 1 for each specific event that they provided (range = 0–4). Parents, who were watching from another room, rated each response as an accurate or inaccurate account of the children’s activities. The total number of plausible events was recorded (range = 0–4). Reliability coding resulted in an interclass correlation coefficient = .93, p < .011, indicating significant agreement.

Zoo (Modified from McColgan and McCormack 2008)

Two trials of McColgan and McCormack’s task were administered. In each trial, an apparatus depicting either a zoo or an aquarium (counterbalanced) was displayed, with five “cages” or “tanks”, and four lockers. In the first trial, there were lockers in front of the first, second, fourth, and fifth cages and, in the second trial, there were lockers in front of all cages except the second. Children were introduced to a cloth doll (“Molly” in the 1st trial, “Jake” in the 2nd), and were told that “Molly/Jake is going to the zoo/aquarium.” Children were given a brief description of a “zoo” and told that “Molly has a camera with her that she wants to bring to the zoo.” The experimenter then explained that Molly must follow the path through the zoo, and that she can only go through the zoo once. Children were asked to point to the start and end of the path (the direction of the path, left to right and right to left, was counterbalanced across participants), and were asked to name all the animals in order. Finally, children were told “Molly wants to take a picture of the (animal) when she goes to the zoo.” The animal selected was always the third animal in the first trial, and the second animal in the second trial. The actual animals in any given position varied because the placement of the animals was random. Children were asked to point to the animal selected to verify that they remembered/understood which it was, and the experimenter pointed out that there was no locker beside that animal’s cage. Children were then told: “You could leave the camera in one of the other lockers and then Molly could get it when she goes to the zoo. Here is a sticker. Put this sticker on the locker that Molly should leave her camera in so she can take a picture of the (animal) when she goes to the zoo.” Finally, children were asked two memory control questions to ensure that they recalled the direction Molly would travel around the path, and which animal she intended to photograph. To receive a score of 1 for each trial, children were required to place the camera in one of the lockers preceding the chosen animal’s cage/tank and answer both memory questions correctly (range 0–2). Reliability coding resulted in an interclass correlation coefficient = 1.0.

ToM Tasks

Diverse Desires (Wellman and Liu 2004)

Children were shown a toy figure of an adult, a picture of a carrot, and a picture of a cookie. The experimenter said “Here’s Mr. Jones. It’s snack time, so, Mr. Jones wants a snack to eat. Here are two different snacks: a carrot and a cookie.” They were then asked the own desire question: “Which snack would you like best?” Whichever snack they chose, they were told that Mr. Jones preferred the opposite snack. They were then asked the target question: “So, now it’s time to eat. Mr. Jones can only choose one snack, just one. Which snack will Mr. Jones choose? A carrot or a cookie?” The order in which the snacks were named was counterbalanced across participants. To receive a score of 1, children had to respond to the target question with the opposite snack from the own-desire question (Total Score: range = 0–1). Reliability coding resulted in a Cohen’s kappa = 1.0.

Diverse Beliefs (Wellman and Liu 2004)

Children were shown a toy figure of a girl and a picture of a house with a garage and a section of bushes. They were told, “Here’s Linda. Linda wants to find her cat. Her cat might be hiding in the bushes, or it might be hiding in the garage.” They were then asked the own-belief question: “Where do you think the cat is? In the bushes or in the garage?” Whichever location they chose, they were told that Linda thought her cat was in the other location. The target question was then asked: “So where will Linda look for her cat? In the bushes or in the garage?” The order in which the options were presented was counterbalanced across participants. To receive a score of 1, children had to answer the target question with the opposite response given to the own-belief question (Total Score: range = 0–1). Reliability coding resulted in Cohen’s kappa = 1.0.

Knowledge Access (Wellman and Liu 2004)

Children were shown a miniature wooden cupboard, and told, “Here’s a cupboard. What do you think is inside the cupboard?” Children could respond with any item, or by indicating that they did not know what was inside the cupboard. Next, the cupboard was opened and the child was shown the content of the cupboard: a small plush dog. The experimenter then closed the cupboard, with the dog concealed inside, and asked the check question: “Okay, what is inside the cupboard?” Children were then introduced to a toy figure of a girl, Polly, and told “Polly has never seen inside this cupboard. Now here comes Polly.” They were then asked the target question, “So, does Polly know what is in the cupboard?”, followed by a memory question, “Did Polly see inside the cupboard?”. To receive a score of 1, the child had to answer “no” to both the target and memory questions (Total Score: range = 0–1). Reliability coding resulted in Cohen’s kappa = 1.0.

Contents False Belief (Wellman and Liu 2004)

Children were shown a clearly identifiable Band-Aid box, and asked “What do you think is inside the Band-Aid box?”. Next, the Band-Aid box was opened to reveal a toy pig, and the experimenter commented “Let’s see… it’s really a pig inside!”. The Band-Aid box was then closed and children were asked the check question: “Okay, what is in the Band-Aid box?” Children were then shown a toy figure of a boy (Peter) and were told, “Peter has never seen inside this Band-Aid box. Now here comes Peter. So, what does Peter think is inside the Band-Aid box?” If children hesitated or did not respond, they were given a forced-choice question, “Band-Aids or a pig?”. Finally, they were asked the memory question: “Did Peter see inside this box?”. To receive a score of 1, children had to respond correctly to both the target (“Band-Aids”) and memory (“no”) questions (Total Score: range = 0–1). Reliability coding resulted in a Cohen’s kappa = 1.0.

Change in Location False Belief (Carlson and Moses 2001)

Children were introduced to a red box and a blue box and to two character puppets: Ernie and Bert. Bert entered the scene and played with a bouncy ball. He then put the ball in one of the boxes (right or left, blue or red, counterbalanced across participants) and left to “go play outside”. Ernie entered, retrieved the ball, and played with it. He then put the ball in the opposite box and left to go “outside”. Bert returned and children were asked the false belief question: “Where does Bert think the ball is?” and the reality control question “Where is the ball really?”. To receive a score of 1, children had to answer both the false belief and reality control questions correctly (Total score: range = 0–1). Reliability coding resulted in a Cohen’s kappa = 1.0.

EF Tasks

Black/White (Inhibitory Control; Simpson and Riggs 2005)

The experimenter showed the child a black card and said, “This card is black right? When you see this card, I don’t want you to say ‘black’. No, I want you to say ‘white’.” The same procedure was repeated with the white card. Children were then given two practice trials in which any errors were corrected. If necessary, up to six practice trials were given until the child could correctly label both a black (“white”) and a white (“black”) card. The 21 test cards were then administered, at a rate of approximately one card every second, with no additional rule reminders provided. The cards were presented in a fixed, pseudorandom order, with the restriction that no more than two cards of one color were presented in a row. Accuracy (number correct out of 21) was recorded (Total score: range = 0–21). Reliability coding resulted in an interclass correlation coefficient = 1.0.

Standard Dimensional Change Card Sort (DCCS) (Cognitive Flexibility/Set-Shifting; Zelazo 2006)

Following the standard version of the DCCS, the two sorting boxes were placed in front of the children, and they were told that they were going to play a card game. The experimenter explained, “This is the color game. In the color game, all the yellow ones go here (points to the tray with the yellow star affixed to the front), and all the red ones go there (points to the box with the red truck affixed to the front)”. The experimenter then demonstrated by sorting a yellow truck card into the yellow star box. Children were asked to sort a red card, and were either praised or corrected for their performance. Children were then asked to sort the six test cards, a mixture of yellow trucks and red stars, in pseudo-random order, with the exception that no more than two of one type were presented in a row. Children were instructed to place them face down into the boxes. Regardless of whether they sorted correctly or not, the experimenter proceeded to the next trial without correction. Each card was labeled according to the relevant attribute as it was handed to the child, e.g., “Here is a red one. Where does this one go?”. The sorted cards remained in the boxes throughout the task. The experimenter then introduced the post-switch phase, explaining, “Now, we are going to play a new game. We’re not going to play the color game anymore, now we are going to play the shape game.” Children were then asked to sort six test cards according to the new criterion (i.e., shape), without any feedback. Again, each card was labeled with the relevant attribute as it was handed to the child (e.g., “Here is a truck. Where does this one go?”). Children’s total score was the total number of correct post-switch card placements they made (Total score: range = 0–6). Reliability coding resulted in an interclass correlation coefficient = 1.0.

Tower of Hanoi (Planning; Carlson et al. 2004)

Children were presented with a small wooden structure with three pegs. Two wooden circles, referred to as “monkeys,” were placed on the pegs, with the smaller one on top. The experimenter also explained the three rules: (1) the monkeys have to stay in the trees, and cannot go into the water (onto the table), (2) only one monkey can jump at a time, and (3) the larger monkey can never go on top of the smaller monkey. A rule check confirmed that children understood all the rules before the task began. The experimenter then produced an identical wooden structure, and explained that the children’s monkeys were “copycat” monkeys and always wanted to look just like the experimenter’s monkeys. Children were asked to copy the experimenter’s pattern using the least possible number of moves. There were six levels of this task, of which the last three involved the addition of a third, smaller disk. The level of difficulty increased with each level of the task, ranging from two wooden disks and two moves to three disks and four moves required to duplicate the pattern. Children were allowed a maximum of two trials of each level and were required to pass one of the trials in order for the task to continue. Relevant rule reminders were provided when a child failed the first trial of a particular level (e.g., “Remember, only one monkey can jump at a time”). Children received a score of 0–6 based on the highest level they achieved (Total score: range = 0–6). Reliability coding resulted in an interclass correlation coefficient = 1.0.

Truck Loading (Planning; Carlson et al. 2004)

In this task, children played the role of a mail carrier, with the task of delivering five color-coded “party invitations” to five matching miniature houses on a miniature street. Children were asked to deliver the “party invitations” to the houses while following the four rules: (1) the street is a one-way street, so the delivery truck can only drive in one direction, (2) the invitations need to be delivered as quickly as possible, so the truck can only drive around the street once, (3) the color of the invitation has to match the color of the house, and (4) the invitations have to be taken off the top of the pile in the back of the truck. Following the warm up, there were four possible levels of the task, ranging from two to five houses. Children were given a maximum of two trials of each level, with relevant rule reminders given for the first failed trial of a particular level (e.g., “Remember, you can only take a letter off the top of the truck”). Self-corrections were allowed only during the loading stage; if they occurred after the child began “driving” the truck, then the trial was scored as a failure. Children received a score of 1 for each level if they correctly loaded the letters into the truck in reverse order. Scores ranged from 0 to 4, corresponding to the highest level achieved (Total score: range = 0–4). Reliability coding resulted in an interclass correlation coefficient = .99, p < .001.

Word Fluency (Generativity; Turner 1999)

To make this task sufficiently engaging for children, we used a stuffed dog, “Scruffles,” to ask children the generativity questions. The experimenter explained that Scruffles was very curious and wanted to know all the animal names they knew. The experimenter then asked children to tell Scruffles as many animal names as they could think of, as fast as they could. After 30 s, the experimenter proceeded to the next trial, in which children were asked to name as many foods as they could in 30 s. The score reflected the number of novel animal names and foods generated within the time limits. Repetitions of previously mentioned animals or foods or vague categories, such as “snacks,” were not included in the score, nor were any foods or animal names mentioned outside of the 30 s. (Total score = total number of items generated in both categories: range = 0–18). Reliability coding resulted in an interclass correlation coefficient = .93, p < .001.

Backward Digit Span (Working Memory; Carlson 2005)

Children were introduced to a doll named Jenny and were told that Jenny says everything the experimenter says, but says it backwards. The experimenter then demonstrated, saying “5–8”, and making Jenny say “8–5”. Children were given two practice trials with feedback, followed by two test trials each with an increasing number of digits, beginning with two digits. The task ended when children erred on both trials of a given level. Children were awarded a score of 1 for each successful trial (Total score: range = 0–5). Reliability coding resulted in an interclass correlation coefficient = 1.0.

Count/Label (Working Memory; Carlson 2005)

In this task, the experimenter showed children three small, 2D wooden objects (e.g., a boat, an apple, and a bird) and children watched while the experimenter labeled them (“boat, apple, bird”). The experimenter then counted the objects out loud (“one, two, three”). Finally, she counted and labeled them each in turn (e.g., “one is a boat, two is an apple, and three is a bird”). Children were then asked to complete all three steps in two test trials using different objects. Children’s ability to correctly count and label was awarded a score of 1 for each trial (Total score: range = 0–2). Reliability coding resulted in an interclass correlation coefficient = 1.0.