Hemispheric asymmetries in the activation and monitoring of memory errors

Previous research on the lateralization of memory errors suggests that the right hemisphere's tendency to produce more memory errors than the left hemisphere reflects hemispheric differences in semantic activation. However, all prior research that has examined the lateralization of memory errors has used self-paced recognition judgments. Because activation occurs early in memory retrieval, with more time to make a decision, other memory processes, like strategic monitoring processes, may affect memory errors. By manipulating the time subjects were given to make memory decisions, this study separated the influence of automatic memory processes (activation) from strategic memory processes (monitoring) on the production of false memories. The results indicated that when retrieval was fast, the right hemisphere produced more memory errors than the left hemisphere. However, when retrieval was slow, the left hemisphere's error-proneness increased compared to the fast retrieval condition, while the right hemisphere's error-proneness remained the same. These results suggest that the right hemisphere's errors are largely due to activation, while the left hemisphere's errors are influenced by both activation and monitoring.     

Source monitoring in Alzheimer's Disease

Source monitoring is the process of making judgments about the origin of memories. There are three categories of source monitoring: reality monitoring (discrimination between self- versus other-generated sources), external monitoring (discrimination between several external sources), and internal monitoring (discrimination between two types of self-generated sources). We investigated whether Alzheimer's Disease (AD) patients, when compared with young and older adults, are impaired at the same level on the three source monitoring categories. We designed three tasks, one for each source monitoring category. In the first task, aimed at reality monitoring, participants had to remember whether objects were previously placed in a bag by themselves or by the experimenter. In the second task, assessing external monitoring, participants had to remember whether the experimenter had previously placed objects in the bag with a black or white gloved hand. In the third task, measuring internal monitoring, participants had to remember whether they had previously placed or imagined themselves placing objects in the bag. Participants showed worse performances in the external and internal monitoring tasks, when compared with reality monitoring. The external monitoring deficit was even more pronounced in AD patients. Regression analyses showed that variation in the external monitoring performances was reliably predicted by inhibition. Our results emphasize the role of inhibitory processes in AD patients' source monitoring decline. The close relation between source and inhibitory decline in AD is interpreted in terms of a common neural base for both concepts.     

Do Programmers Prefer Predictable Expressions in Code?

Source code is a form of human communication, albeit one where the information shared between the programmers reading and writing the code is constrained by the requirement that the code executes correctly. Programming languages are more syntactically constrained than natural languages, but they are also very expressive, allowing a great many different ways to express even very simple computations. Still, code written by developers is highly predictable, and many programming tools have taken advantage of this phenomenon, relying on language model surprisal as a guiding mechanism. While surprisal has been validated as a measure of cognitive load in natural language, its relation to human cognitive processes in code is still poorly understood. In this paper, we explore the relationship between surprisal and programmer preference at a small granularity—do programmers prefer more predictable expressions in code? Using meaning-preserving transformations, we produce equivalent alternatives to developer-written code expressions and run a corpus study on Java and Python projects. In general, language models rate the code expressions developers choose to write as more predictable than these transformed alternatives. Then, we perform two human subject studies asking participants to choose between two equivalent snippets of Java code with different surprisal scores (one original and transformed). We find that programmers do prefer more predictable variants, and that stronger language models like the transformer align more often and more consistently with these preferences.