Outsourcing Memory To External Tools: A Review Of ‘Intention Offloading’ Part 2

Gilbert (2015a) reported three main findings. First, participants were much more likely to set reminders when they had three items to remember rather than just one. Second, participants were more likely to set reminders when they encountered interruptions during the task (a pop-up box asking an arithmetic verification question). 

The relationship between task process and memory is very close. In the process of completing a task, we need to constantly pay attention, concentrate and think. This kind of thinking activity is very beneficial to brain training and memory improvement. At the same time, through continuous practice and completion of tasks, memory will gradually improve.

First, the task process requires a lot of thinking and judgment on our part. We need to pay attention to details, grasp the key points, and clarify our thinking. This kind of thinking and judgment is a complex activity of the brain, which needs to be completed through the conduction and cooperation between the cerebral cortex and various neurons. This kind of brain training strengthens our neural networks and improves our memory.

Secondly, the task process pays attention to details and rules. When completing tasks, we need to follow certain rules and steps. This meticulous behavioral process helps us further improve our memory. These rules and steps require us to establish a logical framework in the brain, making the brain more sensitive and effective in processing various information and details. This is also a kind of memory training.

Finally, the task process requires us to continue to practice and reflect. Through practice and reflection, we can deeply grasp the content and key points of the task, which greatly improves our memory and thinking ability. This kind of reflection and summary repeatedly sorts out and organizes this information and experience in our brains. A deeper impression will remain in our brains and strengthen our memory.

To sum up, the task process has a very important impact on our memory. During the task process, the activities we need to perform such as thinking, judging, following rules, and practicing reflection are all activities that train the brain and improve memory. We should try more tasks, constantly exercise our brains, and improve our memory. It can be seen that we need to improve memory, and Cistanche deserticola can significantly improve memory, because Cistanche deserticola can also regulate the balance of neurotransmitters, such as increasing the levels of acetylcholine and growth factors. These substances are very important for memory and learning. In addition, Meat can also improve blood flow and promote oxygen delivery, which can ensure that the brain receives sufficient nutrients and energy, thereby improving brain vitality and endurance.

Click Know to improve short-term memory

Therefore, the intention of loading was influenced by both the memory load and the nature of the ongoing task in which this memory load was embedded. Third, the conditions associated with greater intention of loading (higher memory load or greater interruption) were also associated with reduced accuracy when participants were forced to use internal memory alone. 

This suggests that individuals set external reminders to mitigate failures of internal memory that might otherwise occur.

An additional purpose of Gilbert (2015a) was to investigate the relationship between participants' performance of the experimental task and their fulfillment of a naturalistic intention embedded within their everyday lives over a longer period. Participants were provided with a unique web link and told that they could earn additional small bonus payments by visiting this link after 2 days, 5 days, and 7 days. 

For those participants who initially reported that they intended to collect all three bonuses, it could then be calculated whether there was a correlation between (a) the number of bonuses they remembered to collect over 1 week, and (b) accuracy on the experimental tasks in the initial testing session. There was a numerically small but statistically highly significant correlation between the performance of the intention offloading task (with a retention interval of a few seconds) and the naturalistic task (with a retention interval of up to a week). 

The intention of adding task had greater predictive validity for the naturalistic task than any of the other measures that were examined, including more traditional event- and time-based prospective memory tasks.

This funding speaks to the question of whether it is appropriate to consider the intention of adding a task as a task of prospective memory. An alternative viewpoint would be that the intention ofoading task is a measure of short-term or working memory, but does not qualify as a task of prospective memory because the retention interval is so short. 

According to this account, the term 'prospective memory' should be reserved for tasks involving a longer retention interval so that participants cannot continuously rehearse their intended action, but instead need to bring it back to mind after a delay (Graf & Uttl, 2001). In our view, this question of terminology is somewhat arbitrary. 

It is undoubtedly the case that the intention offloading task requires participants to form delayed intentions and then fulfill them after a delay. It is also clearly true that the task has a shorter retention interval than typical prospective memory paradigms. Nevertheless, the task has some external validity in the sense that performance predicts participants' fulfillment of a real-world intention over 1 week, with greater predictive validity than more traditional prospective memory tasks. 

This remains the case regardless of whether it is labeled as a prospective memory task or some other sort of task, and none of the conclusions drawn below are affected by this terminological question. To avoid this issue, we generally prefer to use the more neutral term 'memory for delayed intentions' rather than 'prospective memory', which some authors choose to use in a more restricted sense.

Optimality of Intention Ofoading

Setting a reminder involves both a cost (the time and effort of setting it up) and a benefit (the increased chance of remembering). These costs would mount to an unacceptable level if we set reminders for absolutely every activity we intend to perform, including routine daily activities like remembering to go to work, to eat, to sleep, and so on. 

Therefore, individuals need to continually make decisions about whether the benefit of setting a reminder outweighs the cost. While the paradigm used by Gilbert (2015a) allows measurement of how often participants decide to set reminders, it cannot be used to determine how optimal these decisions are. 

To investigate this question, Gilbert et al. (2020) adapted the earlier paradigm so that it can be used to determine whether individuals weigh the costs and benefits of reminders optimally or show a systematic bias either towards or away from using them.

In this paradigm (Fig. 2), participants perform a demanding task, where mean accuracy using internal memory is typically around 50–60%. Alternatively, they can set external reminders, in which case accuracy is typically 90–100%. 

As they perform this task, they are repeatedly offered a choice between (a) using their memory, in which case they earn maximum points for each remembered item, or (b) using reminders, in which case they earn a smaller number of points which varies from trial to trial. Suppose an individual can achieve 55% accuracy using their memory and 100% using reminders. 

Offered a choice between 10 points per item using their memory and 5 points using reminders, the optimal choice is to use internal memory (which would earn on average 5.5 points per item) rather than reminders (which would earn only 5 points). But if 6 points per item are offered using reminders, it becomes optimal to switch to this strategy instead.

In some trials, participants are forced to use either internal memory or reminders. Based on the performance of these trials, the optimal strategy can be calculated for the choice trials. This can then be compared against the actual choice behavior to evaluate whether participants use (a) more reminders than would be optimal, (b) fewer reminders than would be optimal, or (c) the optimal number of reminders. 

Note that the calculation of this measure is individually tailored to each participant's accuracy on the forced internal and external trials. So a particular reminder-setting strategy such as always setting reminders when offered 6 points or higher might reflect an under-use of reminders for a participant with relatively poor memory, but an over-use of reminders for a participant with relatively good memory. 

Studies using this paradigm have consistently found evidence for a systematic bias: Individuals tend to set reminders on a greater number of trials (and, equivalently, use internal memory on a smaller number of trials) than would be optimal (Ball et al., 2022; Engeler & Gilbert, 2020; Gilbert et al., 2020; Kirk et al., 2021; Sachdeva & Gilbert, 2020). 

Individual differences in this bias are stable over time (Gilbert et al., 2020, Experiment 1). Results also show that participants with poorer memory ability tend to set reminders on a greater number of trials than those with better memory ability (Gilbert et al., 2020).

Metacognition as a Trigger for Intention Ofoading

Each time we form an intention, we need to decide whether to remember it internally or set an external reminder. How do individuals make these decisions? One clue comes from the data described above (Gilbert, 2015a), showing that participants are increasingly likely to set reminders in conditions where their performance is poorer (i.e., with a higher memory load or interruptions during task performance). 

This could be explained by the influence of metacognition, our ability to monitor and control our cognitive processes and abilities (Dunlosky & Metcalfe, 2008; Flavell, 1979; Fleming et al., 2012; Koriat, 2007; Metcalfe, 1996; Nelson & Narens, 1990). 

Given that individuals set more reminders in situations where their memory processes tend to be inadequate, this suggests that they rely on a metacognitive evaluation of their memory abilities to trigger the intention ofoading when it is necessary (see also Arango-Muñoz, 2013; Weis & Wiese, 2020).

Previous studies have investigated metacognition of prospective memory with a variety of approaches, including asking participants to make quantitative assessments of their performance (e.g., Cauvin et al., 2018; Meeks et al., 2007; Rummel et al., 2013; Schnitzspahn, Zeintl, et al., 2011) or via questionnaires (Crawford et al., 2006; Rummel et al., 2019). 

Results suggest that individuals do have some metacognitive awareness of their prospective memory abilities (Meeks et al., 2007), but they are often unduly pessimistic, that is, they predict that they will perform worse than they do (see Kuhlmann, 2019, for a review).

However, just because individuals tend to set reminders in conditions when internal memory is more likely to fail, this does not necessarily imply that they do so as a result of metacognitive processes. 

An alternative explanation could be that individuals learn by associative mechanisms the situations where intention offloading is most beneficial, or by earlier instruction to food in particular situations, without directly engaging in metacognitive monitoring of their memory abilities. Direct evidence for a metacognitive influence on intention offloading would therefore require a demonstration that intention offloading is predicted by subjective metacognitive beliefs about memory ability, regardless of the need for offloading as measured by objective memory ability. 

Such evidence was found by Gilbert (Gilbert, 2015b; see also Dunn & Risko, 2016, for a conceptually related effect in a different domain of cognitive offloading). Participants performed an intention of loading task in two phases: first relying on internal memory only, and second with the option of setting reminders if they wished. They also made metacognitive performance evaluations at each phase. 

Results showed that the likelihood of reminder-setting in phase 2 was predicted by objective unaided accuracy in phase 1 – how much they needed reminders, and, independently, by their metacognitive prediction in phase 1 – how much they thought they needed reminders. 

In one experiment (Gilbert, 2015b, Experiment 1a) participants' use of reminders was predicted by metacognitive evaluations, even when those metacognitive evaluations were entirely unrelated to objective accuracy (r = -0.01), providing clear evidence for a metacognitive influence on intention offloading.

In another study (Boldt & Gilbert, 2019), one group of participants was explicitly instructed on how to set reminders if they wished to do so, similar to Gilbert (2015b). The second group was not provided with any instructions, so they could only set reminders if they spontaneously invented the strategy themselves. Results showed that participants in the spontaneous group did indeed spontaneously generate an offloading strategy; however, they did so to a lesser degree than the explicitly instructed group. 

Nevertheless, the same association with confidence was found in both groups: Participants with lower confidence were more likely to set reminders, regardless of objective memory ability. Therefore, even in a situation where the strategy needs to be spontaneously generated rather than explicitly instructed, the intention of loading is still guided by low confidence (see also Hu et al., 2019).

Further evidence for the role of metacognition on intention ofoading comes from a study where participants underwent metacognitive interventions, i.e., interventions designed to influence their metacognitive beliefs (Gilbert et al., 2020, Experiment 3). 

We manipulated the difficulty of practice trials and, independently, the feedback received. This feedback did not deceive participants but described the same level of performance in positive or negative terms, such as "Well done – excellent work! You responded correctly to most of the special circles" versus "Room for improvement. You got some of the special circles wrong". 

Results showed that the metacognitive interventions influenced confidence: Participants were significantly more confident after receiving positive feedback, and when they received easy practice trials. However, there was no effect on objective accuracy. The metacognitive interventions also influenced reminder bias: To the extent that participants became more confident, they relied less on external reminders. 

Further, mediation analysis showed that shifts in reminder bias were mediated by shifts in confidence. Therefore, metacognitive interventions can affect reminder setting without affecting unaided memory performance, providing strong evidence for a metacognitive influence on intention offloading.

Intention Ofoading and the Avoidance of Cognitive Effort

As discussed in Optimality of Intention offloading above, the optimal reminders paradigm typically reveals a positive reminder bias, that is, participants tend to rely on external reminders more than would be optimal. Consistent with a metacognitive account, they also tend to be underconfident, predicting lower accuracy than they achieve (Engeler & Gilbert, 2020; Gilbert et al., 2020). 

In this context a positive reminder bias is rational: Insofar as an individual believes that their internal memory processes are inadequate, they should rely on external resources instead. However, the metacognitive account cannot explain the reminder bias in full. 

In the metacognitive intervention study described above (Gilbert et al., 2020, Experiment 3), one group of participants who received both easy practice trials and positive feedback were significantly over-confident in their memory ability. The reminder bias in this group was significantly reduced, but it was nevertheless still positive, meaning that participants were still more likely to use external reminders than would have been optimal. 

If the reminder bias was wholly attributable to metacognitive bias, then just as an under-confident participant would be expected to over-rely on reminders, an over-confident participant would be expected to show a negative rather than a positive reminder bias. 

This study showed that a positive reminder bias can be observed in the context of both over- and under-confidence, indicating that one or more additional factors must be involved. Sachdeva and Gilbert (2020) provided evidence that one such factor is a preference to avoid cognitive effort.

The concept of cognitive effort is elusive to describe at a mechanistic level (Shenhav et al., 2017). Nevertheless, it is widely argued that cognitive effort is typically aversive (Dreisbach & Fischer, 2015; Kurzban, 2016; Saunders et al., 2017) and individuals generally tend to avoid effortful tasks (Frederick, 2005; Kool et al., 2010). 

Some theoretical accounts have proposed that cognitive effort is a limited, depletable resource that individuals strive to conserve (Baumeister et al., 2007). Although this viewpoint can potentially explain why individuals would avoid stressful tasks, it has encountered serious conceptual and empirical challenges (Hagger et al., 2016; Lurquin & Miyake, 2017). 

Contemporary theories of cognitive effort have focused on an alternative approach, arguing that the feeling of cognitive effort arises from the engagement of relatively domain-general processes that can only be deployed for a limited number of simultaneous tasks (Kurzban et al., 2013). Such domain-general processes incur an opportunity cost. In other words, exercising cognitive effort on one activity precludes its use on another. An aversion to cognitive effort can then be understood as a drive to reduce this opportunity cost.

This model can potentially explain both why individuals might avoid effortful tasks, and why remembering an intention with internal memory may be more effortful than using external reminders. There are well-known limits to the quantity of information that can be actively maintained in short-term or working memory (Bays & Husain, 2008; Miller, 1956). 

Maintaining one active intention, therefore, incurs an opportunity cost, since this may preclude simultaneously maintaining another one. By contrast, external tools such as smartphone alerts have an effectively unlimited capacity. As a result, individuals may prefer external reminders over internal memory because they incur a lower opportunity cost.

Sachdeva and Gilbert (2020) argued that manipulating performance-based financial rewards allows a test of whether effort avoidance influences intention offloading. If excessive use of reminders is explained by metacognitive error alone, then it should not matter whether or not performance is incentivized with financial reward. 

Regardless of the reward, participants would be selecting the optimal strategy, based on their metacognitive beliefs. But if effort avoidance makes an additional contribution, then the reminder bias should be reduced by financial incentives. Seeing as individuals are more likely to exert cognitive effort when they have a financial incentive to do so (Aarts et al., 2010; Padmala & Pessoa, 2011), we would predict that performance-based rewards would at least partially overcome the preference to avoid cognitive effort, and therefore the reminder bias would be reduced. 

In other words, a reduction of the reminder bias as a result of financial reward would be diagnostic of an influence of effort-avoidance, rather than a metacognitive error. This is exactly what was found by Sachdeva and Gilbert (2020): the bias towards external reminders was significantly reduced, but not eliminated, when participants received a financial reward for their performance of the task. 

Therefore, a second factor contributing to individuals' intention offloading decisions, in addition to metacognitive belief, is a preference to avoid cognitive effort.

Strategy Perseveration

A third factor contributing to the intention of loading was demonstrated by Scarampi and Gilbert (2020, Experiment 2). Participants performed an intention of loading task in two phases. In the first phase, they were either forced to use reminders or forced to use their memory. In the second phase, they had a free choice about whether to set reminders. Results showed that despite their free choice of strategy, participants tended to perseverate with whichever strategy they had used in phase 1. 

Therefore, previous instruction or experience in adding strategy is an additional factor that influences ongoing offloading behavior. Scarampi and Gilbert (2020, Experiment 1) also found that previous use of an offloading strategy did not influence subsequent unaided memory, at least in the short term.

These findings are relevant to the debate about the potential long-term benefits or harms of cognitive technology. It has been argued at least since the time of Socrates that relying on external tools rather than brain-based processes might lead to a harmful decline in cognitive ability (Kallick, 1989). This fear has found expression in contemporary debates about whether technologies such as Google are 'making us stupid' (Carr, 2008) or leading to 'digital dementia' (Moledina & Khoja, 2018). 

We consider that these fears are overstated because they hypothesize long-term harms based only on short-term evidence, and they disregard evidence showing that offloading can have positive as well as negative cognitive impacts (Cecutti et al., 2021; Runge et al., 2019; Storm & Stone, 2014).

Furthermore, the fear that using cognitive tools will have harmful consequences presents only one side of the cost-benefit calculation. The other side is that failing to use cognitive tools may be harmful if such tools are available and useful. 

Evidence shows that external reminders are highly effective. For example, participants in Gilbert et al. (2020) had a forgetting rate of about 45% when using their memory but only around 5% using external reminders; in other words, using reminders reduced the forgetting rate by almost an order of magnitude (see also Jones et al., 2021). 

In this context, an individual who decides against using reminders, for example, an older adult who believes in the importance of 'use it or lose it' to maintain cognitive health, will be depriving themselves of an extremely effective and convenient tool, which can promote all the benefits that come with being able to fulfill one's delayed intentions effectively. 

In addition, the strategy perseveration effect reported by Scarampi and Gilbert (2020) suggests that an earlier decision to forgo reminders can influence future strategy as well. This could have a particularly harmful impact in the context of cognitive decline, which makes the effective use of compensatory tools even more important.

Intention Ofoading across the Lifespan: Child Development

When and how do children develop the ability to supplement their brain-based cognitive processes with external reminders? Redshaw et al. (2018) investigated this question in children aged approximately 7–13 years, using an intention of loading task similar to that of Gilbert (2015a) administered with a touchscreen tablet computer. 

As in Gilbert (2015b), the task was performed in two phases, first using unaided memory and second with the option to set reminders. There were two levels of difficulty (one item vs. three items to 67 1 3 Psychonomic Bulletin & Review (2023) 30:60–76 remember) and participants provided separate metacognitive predictions for each level of difficulty in each phase.

Older children (11+ years of age) offloaded strategically in the same way as adults (Gilbert, 2015a): they were much more likely to set reminders when they had more items to remember. However younger children (< 9 years) were equally likely to have intentions, regardless of the memory load. 

Could this be explained by a lack of metacognitive knowledge in the younger children? If they failed to understand the increased likelihood of forgetting at the higher memory load, it would be unsurprising if they did not compensate for this with increased reminder-setting. However, results from the metacognitive judgments ruled this out: if anything, the younger children were more sensitive than the older children to the increased likelihood of forgetting at the higher memory load. 

Therefore, it seems that the younger children possessed the metacognitive knowledge that they were more likely to forget at the higher memory load, but they lacked the metacognitive control to translate this into strategic reminder-setting targeted at the more difficult trials. 

This distinction between metacognitive knowledge and control is consistent with prior evidence from the metamemory literature – for example, younger children can distinguish between easy and difficult items for a memory test, but only older children allocate more study time to the more difficult items (Dufresne & Kobasigawa, 1989; Lockl & Schneider, 2004; Masur et al., 1973).

For more information:1950477648nn@gmail.com