The Internal Validity Of Intrinsic Frame Of Reference

One important aspect of our daily activities is to navigate in various complicated environment, and an important step for accurate navigation is determining the appropriate position and orientation of the destination. Since Tolman asserted that rats have cognitive maps of the spatial layout of their environments, spatial representation has been well studied and discussed. Reference systems in spatial processing can be classified into two types, those egocentric reference system which specifies a location with respect to the observer, and environmental reference system which specifies a location with other locations independent of the observer. McNamara and his colleagues' proposed a different issue that the locations of objects in memory are specified with intrinsic reference systems, which is a type of geocentric reference systems. In intrinsic reference frames, the location of an object is defined with respect to other objects in the layout. The objects in environment can form infinite number of possible intrinsic axes, but because of the perceptual grouping principles such as proximity and similarity and egocentric experience, some axes will be more salient than others.Although McNamara and his colleagues have accomplished a large volume of valuable works on intrinsic reference systems, and most of their researches approved their hypothesis of intrinsic frame of reference, however, their conclusion of intrinsic frame of reference is based on a very limited and special set of examples; the conclusion could be based on the artifact of the layout or the specific experimental method they used. Firstly, in most of McNamara's research, the objects are positioned in very regular way; much like on a chest game board, which makes the intrinsic structures of object-to-object the most salient reference axes. Furthermore, there are only a very limited numbers of angles between objects (e.g.,±tg1,±tg1/2), thus subjects would remember the layout and the relation of the objects very easily. Hence, further researches are required to test whether intrinsic reference frames are effective in other irregular and inartificial environment. Secondly, in Mou and McNamara's research, participants viewed the layout from the 315°position, and were instructed to learn learning the locations of objects in a room according to egocentric 315°-135°axes or according to non-egocentric 0°-180°axes. The instructional manipulation was accomplished by pointing out that the layout could be seem in columns appropriate to the learning axis, and required participants memorize the layout column by column, consistent with the appropriate axis condition. Mou and McNamara concluded that the different results pattern for egocentric 315°-135°group and non-egocentric 0°-180°group in their research was produced by the intrinsic structure of the layout, they ignored that it would be more reasonable to conjecture that instruction words or experimental methods play an important role in spatial memory. Considering these possibilities, further researches were needed to separate the effect of intrinsic structures to instruction words and experimental manipulations.The goal of the research described in this paper was to determine in what conditions the intrinsic frames of reference works. In experiment 1, seven objects were used and all of them were positioned regularly in a rectangular room. Participants learned the layout of objects from a single view (0°or 135°) and then made judgment of relative direction using their memories of the layout. In experiment 2, one object was added and positioned irregularly, in half of the trials, participants were required to judge the relative direction of the regularly positioned objects and in other trials they were required to judge the relative direction of the oddly positioned object. For participants who learned the layout from 0°position. In experiment 3, the spatial layout and procedure was the same with experiment 2, except that all participants viewed the layout from 45°position and were instructed to learn the layout from egocentric 45°-225°axis or non-egocentric 0°-180°axis.The specific findings are presented as follows: Firstly, participants' learning experiences play an important role in spatial memory, when participants learned the layout from single viewpoint; they inclined to represent the layout in egocentrically, with experienced heading performed quicker and more accurate than novel headings, no matter the experienced heading was aligned or misaligned with the geometric frame of environment. Secondly, when participants were asked to point to one object repeatedly from different headings, they inclined to form an orientation-free representation of that object. In both experiment 2 and experiment 3, the results pattern for the oddly positioned object seemed to be different with results for other objects, with main effect of imagined heading and interaction of imagined heading and learning axes not significant, and no sawtooth pattern was evident. Thirdly, the intrinsic reference frame could benefit the representation of spatial layout, but this advantage depend on the highly regularity of spatial layout and participants' learning experience. In experiment 1, all of the seven objects were positioned regularly and participants learned the layout from single viewpoint which (0°) was aligned or misaligned (135°) with the walls of the room and the intrinsic structures of the objects. The results of 0°learning condition showed a sawtooth pattern across novel headings, with novel headings aligned the geometric frame of environment and intrinsic frame of objects performed better than other novel headings. However, in experiment 2, when one object was added and positioned oddly, although the positions and test trials for regularly positioned objects were exactly the same with experiment 1, the sawtooth was no longer evident, even for participants who learned the layout from 0°orientation, which was aligned with the walls of the room and the intrinsic axes formed by the objects. Fourth, the importance of intrinsic reference frames in some extent could be promoted by experimenters' instructions and requirements. In experiment 3, the spatial layout and experimental methods were similar with experiment 2, but some of the intrinsic axes were pointed out and emphasized by experimenter. The results indicated that participants could represent the layout according to given intrinsic axis as the experimenter instructed and required to, with non-egocentric 0°-180°learning condition showed a sawtooth pattern across novel headings for regularly positioned objects. But no sawtooth was evident for the oddly positioned object, indicated that the intrinsic frames of reference didn't benefit the representation of the oddly positioned object even participants were instructed and required to remember the object positions according to the intrinsic structure of the layout. Finally, in both experiment 2 and experiment 3, the reaction time for regularly positioned objects was longer than the oddly positioned object, but the angular error showed no difference. This results pattern may be attributed to the testing procedures, half of the test trials in experiment 2 and experiment 3, participants need to point to only oddly positioned object and in other trials they need to point to the seven regularly positioned objects. Which means that the oddly positioned object was retrieved and pointed 32 times and each regularly positioned objects was retrieved and pointed about 4.6 times for each participant in testing phrase. The repeatedly retrieval and pointing made the oddly positioned object more familiar and easier to locate than other objects. On the other hand, when participants were required to point to regularly positioned objects, the intrinsic structure of the layout provided extra cues to improve the pointing accuracy. As the distances between regularly objects were the same (1m) and the angles between objects were limited (e.g.,±tg1,±tg1/2), subjects could figure out the exact or approximate angles to improve the accuracy when they were required to judge the relative direction of objects, however, this improvement need time to achieve. According to these results, the intrinsic structure of objects' position possibly benefited the judgment of relative direction, but did not benefit the representation of spatial layout.In summary, these experiments indicated that under some circumstances, intrinsic structure of the layout could benefit the spatial memory, but the benefit depends on the regularity of the object-locations, and the experimental manipulations which focus participants' attention to the intrinsic structure of layout in learning phrase. What's more, the findings in present research suggest that the intrinsic structure of spatial layout possibly provided extra cues for participants to judge the relative directions of objects, rather than provided reference frames in spatial representation. This research also implies that the representation of room-sized layout could be allocentric if participants are sufficient familiar with the locations of objects. For example, when participants were asked to point one object from different headings, they became familiar with this object and constructed an enduring allocentric representation of it.