Many cognitive neuroscience paradigms designed to assess relational encoding and retrieval suffer from a problem that limits their application to clinical populations - they leave the nature of the encoding strategy up to the individual, leaving open the possibility that poor task performance simply reflects a failure to apply relational processing, rather than a fundamental deficit in the ability to engage in such processing. Ranganath and colleagues developed a task to investigate the cognitive and neural mechanisms underlying relational encoding and retrieval (Murray & Ranganath, 2007) (Blumenfeld & Ranganath, 2006) that explicitly controls whether participants engage in item-specific or relational memory processing. This task has been studied with verbal materials using functional magnetic resonance imaging (fMRI) to specifically link memory success following relational versus item-specific encoding to increased engagement of the dorsolateral prefrontal cortex. This paradigm does an excellent job meeting the CNTRICS selection criteria of strong construct validity, a specific neural mechanism, and appropriateness for functional imaging. However, this paradigm also has a number of features that are suboptimal for high throughput clinical trials research, including; prominent task switching demands, high meta-cognitive demands, and use of English word stimuli. Therefore, the RISE was designed to modify test stimuli and experimental procedures of the Murray and Ranganath procedure (2007) to optimize patient tolerability (fatigue and task comprehension) while maintaining strong construct validity to allow us to test the prediction that individuals with schizophrenia will have a differential deficit in their ability to engage relational versus item-specific encoding and retrieval processes to support episodic memory. To accomplish these goals, the RISE was developed with visual object rather than word stimuli, requires participants to alternate encoding tasks between blocks rather than individual trials, and uses confidence ratings and a receiver operator characteristics analysis rather than a remember/know procedure to reduce meta-cognitive demands and obtain independent estimates of familiarity versus recollection based retrieval.
We have developed three parallel versions of this task. The stimuli consist of 432 visual object representations of the word stimuli employed in the original study. Visual objects rather than words were used to improve patient understanding of encoding instructions and facilitate use in future non-English speaking clinical trials settings. Visual objects were selected from a standardized corpus of 2,500 color photographs (http://cvcl.mit.edu/MM/), and divided into three groups of 144 stimuli for each of the three test forms. For each test form, a list of 36 items was assigned to item encoding, 36 were assigned to relational encoding, and 72 served as foils for recognition testing. Lists and test forms were matched for word length and frequency based upon original study norms. There are three components to each version:
- Item-Specific Encoding - 36 stimuli are presented for two seconds each, with a one second inter-stimulus-interval (ISI), and subjects make a two-button "yes/no" response to indicate whether objects were "living". This levels-of-processing manipulation controls for potential group differences in strategy generation (Ragland, et al., 2003), and was superior to a "pleasant/unpleasant" decision that produced ceiling effects in pilot subjects.
- Relational Encoding - 18 pairs of visual objects are presented for 4 seconds each, with a one second ISI, and subjects make a two-button "yes/no" response to indicate whether one item could fit inside the other item.
After encoding, two retrieval tasks are administered:
- Item Recognition - all 72 objects studied during encoding trials (36 item-specific and 36 relational targets) are randomly intermixed and presented with 72 new unstudied foils, and subjects indicate whether each test item was "old" (left hand response) or "new" (right hand response) and rated their confidence using one of three buttons (i.e., 3=high, 2=medium, 1=low). Confidence ratings rather than the original R/K procedure were used given previously noted concerns about the meta-memory demands of R/K.
- Associative Recognition - all 18 original object pairs studied during relational encoding are randomly intermixed and presented with 18 rearranged object pairs consisting of items presented during different encoding trials and not originally paired together. Subjects make a two-button "yes/no" response to indicate whether items in each pair had been presented "together". To prevent additional encoding of the relational object pairs, the item recognition task precedes the associative recognition task. Both retrieval tasks are self-paced and subjects are instructed to "work as quickly and accurately as you can." Subjects are required to successfully complete practice tasks prior to the start of encoding and retrieval phases to ensure comprehension. These practice components are included in the Eprime files, along with instructions. The complete procedure, including practice, encoding and retrieval requires 15-20 minutes to complete.
These versions are currently being used in our test-retest reliability, and could change based on the results of that study.