Where she looks different from the last

slice: distributed cognition in the activity of fMRI motion correction

Morana Alac

Cognitive Science

UCSD

The present analysis is a part of a broader ethnographic study on the production of visual fMRI evidence, theoretically based in Distributed Cognition (Hutchins, 1995) and Conceptual Integration theory (Fauconnier & Turner, 2002). The study involves observations of scientific practices conducted over a period of 6 months in three laboratories in the Cognitive Science department at UCSD and the Salk Institute. Data collection methods include direct observation, video, semi-structured interviews, and analysis of documents such as journal articles, laboratory manuals and scientific correspondence.

In this presentation, I analyze video recordings of a specific scientific practice, i.e. an early stage in the preparation of the fMRI data for further analysis. During this stage, among other types of manipulation, the data have to be corrected for motion. Motion correction transforms each scan (volume or image acquired every few seconds) by using rotation and translation, so that the image of the brain within each scan is aligned with that in every other scan. I analyze five excerpts from one 35-minute session in which an expert explains to a novice how to asses the extent of head motion. This is done by viewing brain images captured at different moments in time. In the course of motion correction the scientists can decide whether or not to reject a data set.

Analysis suggest that the situated action in these excerpts generates complex conceptual structures where initial conceptual entities are compressed. Through the mechanism of conceptual compression two different conceptual entities are considered to be an integrated item (Fauconnier and Turner, 2002). In this way, the physical person in the scanner and the digital representation of her brain slice are conceived as if a single item. Thanks to this conceptual mechanism, the abstract representations become embedded in real world situations of actions. This process of conceptual manipulation can be characterized by what Fauconnier and Turner (2002) call "blends on human scale". Blends, hybrid cognitive models, are defined to be on human scale with respect to the ease with which they are produced, manipulated, or remembered; they lead to greater efficacy in comprehension, greater memory for complex structure, and are associated with the production of creative solutions. The analysis illustrates how the embodied and distributed nature of the task influences complex conceptual structure of the scientists, and hence produces greater cognitive efficacy.

In contrast to the very impersonal style of scientific papers, as well as the use of fMRI data for very general claims about human cognition, at this stage of data analysis the images are often used to speak about a single subject. Through metonymical mappings, representations of the brain slices stand for a particular person.

For example, in order to understand the expert utterance:

1 Where she actually makes the movement =

2 = where she looks different from the last slice?

the interlocutors have to metonymically link the subject in the scanner with the subject’s brain (i.e., the brain stands for the person). In this example, she in line 1 refers to the subject in the scanner, while she in line 2 refers to a single image of that subject’s brain. Thus, the digital representation of the brain slice at a particular moment in time has to be conceptually linked with brain of the person in the scanner: the picture of one slice at a particular moment in time stands metonymically for the brain and represents the brain, and hence it stands for the person in the scanner. The subjects are looking at the digital images on the computer screen and are able to infer the movement of the subject's head in the scanner. This is achieved by constructing a powerful conceptual compression which the expert implicitly communicates to the novice. The head movement that caused a particular effect in the digital representations (i.e., the two images taken at different points in time are not aligned with each other) is compressed with that effect to one conceptual entity, representing at the same time the person in the scanner and the movement detected by observing the non alignment of the images at the computer screen. This composed mental entity is manipulated through this interaction.

In addition to the linguistic utterance, in line 1 the expert is using both her arms to construe the meaning through gesture: she swings her right arm upward while the left remains parallel to the ground. (motion resembling scissors opening). Her right arm, swinging vertically upwards, can be interpreted as representing the subject's head moving up, while the left arm stays still, marking the position from which the movement initiated. However, this gesture can also be seen as representing the blend between the brain slice representation and the subject moving. In this case, the right arm is representing the slice representation/subject moving up, while the left arm maps onto the slice representation that serves as an indicator with respect to which the movement is detected. The fact that interlocutors are manipulating a complex conceptual construction is also evident from line 2. The expert implies the existence of a conceptual entity which has some properties of the subject in the scanner and others of the image of the brain slice at the computer screen. The manipulation of this entity is achieved through the use of language, hand movements and images on the computer screen.

The process of motion correction is conducted by the use of different semiotic resources that enable the interlocutors to "see" what the subject is doing and how this is reflected in the data set. The materiality of the world, as well as bodily movements and gestures of both interlocutors are manipulated and used to create the distributed representations (Hutchins, 1995) that aid understanding and goal achievement.

WORK CITED

Fauconnier, G. & Turner, M.

2002 The Way We Think: Conceptual Blending and the Mind's Hidden Complexities, Basic Books.

Hutchins, E.

1995 Cognition in the Wild. Cambridge, MA: MIT Press.