Mnemonic Similarity Task (MST)
AKA: Behavioral Pattern Separation Task - BPSO
Background
The Mnemonic Similarity Task (MST), previously known as the 'Behavioral Pattern Separation Task', is a sensitive measure for memory changes across the life span and is useful for the early detection of memory impairments. It was developed by Shauna M. Stark and colleagues in 2013 to investigate people's ability to organize overlapping inputs, e.g. similar images, into distinct memory representations as a marker for early changes in memory impairments.
In their study they used a two-stage memory paradigm. During the initial (memory encoding) stage, they presented people of various ages with a set of images depicting a variety of objects and asked their participants to perform a basic categorization task (e.g. 'is that an indoor or an outdoor object?'). The second stage of the task consisted of a (surprise) recognition task. This recognition task not only included the 'old' items previously seen and some very different 'foil' objects - as is done in any 'regular' recognition test', but crucially it also included a set of 'lure' items that depicted similar objects of various levels of the previously seen items. By analyzing participants' performance on their classifications of these lure items, Stark et al showed that the Lure Discrimination Index (LDI) was a highly sensitive marker for early mild cognitive impairments (MCI) in a group of older adults, even if recognition memory of the actual 'old' items itself was still unremarkable compared with age-related peers.
Stark et al refer to the process of generating distinct memory representations of overlapping instances as 'pattern separation' and originally named the MST as the 'Behavioral Pattern Separation Task - Object Version' (BPS-O). They renamed the task in 2015 to more accurately reflect the nature of the task as a behavioral memory assessment tool based on image similarity without making inferences about underlying physiological processes such as pattern-separation.
While the MST shows to be a highly sensitive instrument to assess early memory impairments in academic settings, its administration proves difficult for clinical purposes or large-scale testing. The procedure of the MST was thus revised by Stark and colleagues in 2023 and has become known as the Optimized Mnemonic Similarity Task (oMST).
The Millisecond implementation of the MST provides four different stimuli sets.
Task Procedure
At the beginning of the MST procedure, the set of 192 stimuli is randomly divided into 64 targets ('old'), 64 foils ('new') and 64 lures ('similar'). The lure items are selected first with the constraint that each of the five lure categories (ranking the similarities of the lure item to the target on a scale from 1-5) contains roughly the same number of items.
During the encoding phase (Stage 1), participants are presented with a random order of the 64 targets items and 64 items that are similar to the selected lure items (128 trials total). Each image is presented for 2 seconds total and participants have to decide whether the item is an 'outdoor or an indoor object': If it's an outdoor item, they press "O", if it's an indoor item they press "I" (though actual performance is irrelevant and is not further analyzed).
The recognition phase (Stage 2) starts out with an instruction video explaining the procedure of the recognition phase as well as the three response keys (V,B,N for 'old', 'similar', and 'new', respectively). The 64 target items (previously seen), the 64 foil items (never seen) and the 64 lure items (similar to the ones seen in stage 1) are then presented in random order. Each image is presented for 2 seconds and participants have to decide if this was a new, an old, or a similar item.
What it Measures
The MST is a tool to measure episodic memory impairments
Psychological domains
- Episodic Memory: Ability to remember a specific personal event in time.
- Recognition Memory: Capability to correctly identify previously encountered stimuli among unfamiliar distractors.
- Visual Long-Term Memory: Durability and storage capacity of high-detail visual representations over time
- Object Recognition: Lower-level visual processing required to categorize what an object is
- Sustained Attention: Ability to maintain focus across a continuous stream of dozens of visual stimuli to catch subtle differences.
Main Performance Metrics
- LDI: Lure Discrimination Index, comparison metric of proportion correctly identified Lures and foils that have been incorrectly categorized as Lures; Main MST measure (also referred to as Lure Discrimination Index
- TRI: Traditional Recognition Index, comparison metric of proportion correctly identified Targets and foils that have been incorrectly categorized as Targets
Psychiatric Conditions
MST performance has been found to be impaired in the following patient groups
- Mild Cognitive Impairment (MCI)
- Asymptomatic/Preclinical AD Risk Groups
- Alzheimer’s Disease (AD)
- Parkinson’s Disease (PD)
- Major Depressive Disorder (MDD)
- Schizophrenia
Test Variations
The Mnemonic Similarity Task (MST) is an object recognition memory task originally called Behavioral Pattern Separation Task (Stark et al, 2013) and renamed in 2015. It can be used to study memory performance over the life-span. This script uses stimuli set C.
This MST script uses stimuli set D.
This MST script uses stimuli set D.
This MST script uses stimuli set F.
The oMST is a continous form of the classic Mnemonic Similarity Task (MST), an object recognition test sensitive to changes in memory performance. Whereas the MST uses a two phase testing paradigm, the oMST combines both phases into a single one (thus it's also referred to as the continuous MST), cutting down on testing time. This script provides the oMST run with stimuli set 1 (3 different trial orders provided)
This script runs the oMST with stimuli set 2 (3 different trial orders provided)
This script runs the oMST with stimuli set 3 (3 different trial orders provided)
This script runs the oMST with stimuli set 4 (3 different trial orders provided)
This script runs the oMST with stimuli set 5 (3 different trial orders provided)
This script runs the oMST with stimuli set 6 (3 different trial orders provided)
References
Search Google Scholar for peer-reviewed, published research using the Inquisit Mnemonic Similarity Task (MST).
Kirwan, C. B., & Stark, C. E. (January 01, 2007). Overcoming interference: an fMRI investigation of pattern separation in the medial temporal lobe. Learning & Memory (cold Spring Harbor, N.y.), 14, 9, 625-33.
Stark, S.M, Yassa, M.A., Lacy, J.W. & Stark, C.E.L. (2013). A task to assess behavioral pattern separation (BPS) in humans: Data from healthy aging and mild cognitive impairment. Neuropsychologia, 51, 2442–2449.
Kim, J., & Yassa, M. A. (January 01, 2013). Assessing recollection and familiarity of similar lures in a behavioral pattern separation task. Hippocampus, 23, 4, 287-94.
Ally, B. A., Hussey, E. P., Ko, P. C., & Molitor, R. J. (January 01, 2013). Pattern separation and pattern completion in Alzheimer's disease: evidence of rapid forgetting in amnestic mild cognitive impairment. Hippocampus, 23, 12, 1246-58.
Stark CEL, Noche JA, Ebersberger JR, Mayer L and Stark SM (2023) Optimizing the mnemonic similarity task for efficient, widespread use. Front. Behav. Neurosci. 17:1080366. doi: 10.3389/fnbeh.2023.1080366