Experiments | Context
Context refers to the broadest category to which an experiment belongs to within the realm of brain imaging, also called the main effect or interaction effect.
The available categories for context are:
- Normal Mapping
- Experimental Design
In the functional database, by far the most common context is normal mapping which is around two thirds of the database. The second most used context is disease with 20% and then pharmacology with 5% of the experiments. However, in the structural database, disease is the majority with over 80%, followed by normal mapping at 10% of the VBM experiments.
While multiple contexts can be applied to an experiment, this is rare, with over 90% of the experiments in each database using only one. An example of an exception is when the age of the patient is closely examined as a factor in the progression of a disease, both disease and aging apply. Similarly, when coding an experiment investigating female vs male normal subjects, both gender and normal mapping should be chosen.
In regards to the BrainMap database, “normals” are a somewhat fluid concept. Subjects that may be considered normal in one experiment might not in another. For example, in some situations a cigarette smoker without a formal diagnosis could be used as a healthy control, however when grouped with exclusively smoking people should be classified as within pharmacology. The same could be said for overweight subjects and the context of disease.
Normal Mapping is appropriate when the experiment exclusively contrasts imaging data within exclusively healthy or normal subject groups. The context of normal mapping also excludes any subjects under the influence of drugs or undergoing any treatments. So normal mapping experiments should not be co-coded with disease, pharmacology or treatment. For experiments that have a borderline subject group that looks like it could use both normal mapping and one of these contexts, do not code it as normal mapping.
In some cases, normal mapping can be co-coded with other contexts. Children and elderly normals, non-native speakers, and groups of exclusively left-handed or female subjects, are all eligible to be considered Normal Mapping, provided they don't have another reason to be excluded.
We consider allowing subjects from other contexts into the normal mapping context an acceptable compromise between including all available data and keeping the data quality high. Because of this, we have set "Context is Normal Mapping" as a default search criteria in Sleuth and also use it when performing large scale data-mining. If your analysis should be more stringent, try additional criteria such as "Context is NOT Genetic" or "Context is NOT Aging". Overall, the best way to ensure you have the most wanted and representative sample is a careful review of the search results for eligibility.
The context of disease is used for experiments which are attempting to isolate the effect of a physical or mental illness. This is achieved by investigating a patient group with a relevant diagnosis. This group can be studied in isolation, parsing out the heterogeneity of the disease, or in contrast to a group of healthy controls. If one of the subject groups used in an experiment have a diagnosis, then the context should probably include disease.
The term disease is used broadly, including developmental, genetic and acquired disorders of any etiology. For example:
- (Functional) A study looks at fibromyalgia patients and their brain activity when perceiving pain.
- (Structural) A researcher finds grey matter density reduction in women with anorexia nervosa.
Pharmacology is an appropriate selection for context when a particular drug is administered before the patient is scanned or when a patient regularly adheres to a medication treatment regimen. Please note that "Pharmacology" should not be co-coded with Normal Mapping. Using this context requires the experiments appropriate Pharmacology Class to be defined in order to save the experiment in Scribe. An example from Sleuth, would be the following: After being administered either ketamine or a saline solution intravenously subjects were asked to perform a facial emotion recognition task (Abel, 2003)
Treatment is selected as a context when activation differences are expected to be due to treatment or therapy given to the subjects. An example from the functional database is the De Nil study where researchers looked at the short and long-term effects of a treatment for stutterers (De Nil, 2003). This context should not be co-coded with Normal Mapping.
Aging is chosen as a context when two subject groups of different ages are compared (i.e., verbal fluency in adults and children). These can be two healthy groups who only differ in age, but can also be groups of subjects who share the same diagnosis and thus allow a closer look at the progression of diseases like Alzheimer’s marked by cognitive decline with age.
Emotion is often selected as a context when the paradigm uses emotionally valenced stimuli. For example, a study that looked at subsequent memory of negative vs. positive pictures would be coded as a context of “Emotion”.
Experimental Design is only for studies in which a technical aspect of the paradigm is being manipulated. For example, when comparing the effect of the rate of presentation of the stimuli (“Self-Paced n-back vs. Fixed-Paced n-back” or “Slow Words vs. Fast Words”).
Gender is a context chosen for experiments that focus on gender differences. An example would be mental rotation task performed male and female subjects. Remember, that subjects can be all male, all female or both. As long as the focus is on gender, it should be coded as such.
Genetic would be an appropriate selection when a subject/patient group’s condition involves a genetic anomaly. Remember, most often times these are rare and while many diseases are associated with a genetic component, genetic disorders may/may not be heritable. Down syndrome, also called Trisomy 21, is one of the most common, non-inherited genetic syndromes.
Some examples are the following:
- A study which enrolls subjects with Monoamine oxidase (MAO-A) genetic variations and explores inhibitory control and areas associated with it.
- A study which examines the structure of older adults’ brains at genetic risk for Alzheimer’s disease.
Handedness is a context chosen for experiments that focus on differences in handedness (i.e., left vs., right, dominant vs. non-dominant). For example Tzourio-Mazoyer's 1998 study that looked at speech comprehension in left-handers vs. right-handers.
Learning is often chosen as a context when studies focus on a specific learning pattern, e.g. learning a sequence, motor/visuomotor learning, sensory learning or when subjects undergo training for particular task. An example might be the following: a study investigates persons with Attention-deficit/hyperactivity disorder (ADHD) learn. In the contrast, brain activation and task performance of ADHD learners are compared before and after the training program.
Language is an appropriate context selection when a study focuses on a particular language spoken; primary vs. secondary language; cross-linguistic or the comparison of different languages; or when subjects are multilingual. An example from Sleuth is the Klein’s (2001) cross-linguistic study that included Mandarin Chinese and English speakers and looked at tone perception.