The question joins two things that have been studied apart
The question has two parts. The first is dimensionality: how many independent axes a learned mental space has. The second is reference frame: whether a person locates things relative to their own body (self-centered) or relative to the surrounding world (world-centered). Answering the question requires a study that measures both in the same people and checks whether one predicts the other.
The closest work on dimensionality comes from physiotherapy students learning gross anatomy. That study used data-modeling methods, including multidimensional scaling and Pathfinder networks, to describe the structure of the students' cognitive structures, which is a way of capturing how many dimensions organize what they learned [4]. But it does not look at self-centered or world-centered reference frames at all, so it cannot say anything about the second half of the question [4].
On the other side, discussions of self-centered and world-centered viewpoints appear in work on experience and on personal growth. One paper frames the contrast as a difference between self-centered and world-centered viewpoints [2], and another describes a shift from a self-centered to a world-centered outlook as part of lifelong learning [3]. Neither measures the dimensionality of a learned space, so neither can test the link.
Why the link is plausible but still unproven
It is easy to see why someone might expect dimensionality to matter. A richer, higher-dimensional mental space could in principle support more ways of locating things, including locations defined by the world rather than by one's own body. But that is a hypothesis, not a finding. None of the available work tests it, and the paper that measures dimensionality explicitly does not address reference frames [4].
The reference-frame papers likewise describe viewpoints without tying them to any measured property of a learned space [2][3]. One even treats the move toward a world-centered outlook as a matter of awareness across dimensions of perception, emotion, and life [3], which is a different use of the word "dimensions" than the structural sense used in the cognitive-modeling study [4]. That difference in meaning is a real obstacle: the two lines of work are not even using "dimension" the same way.
What a real test would need to measure
To answer the question, a study would need to do three things at once: estimate the dimensionality of each person's learned space, independently classify whether they rely on self-centered or world-centered cues, and then check whether the first predicts the second. The anatomy study shows the dimensionality measurement is feasible [4], and the viewpoint discussions supply the conceptual contrast to classify [2][3]. What is missing is a study that puts them together.
Until then, the practical answer for a curious reader is to treat the proposed link as an open hypothesis. The evidence is consistent with the idea that learned structures have measurable dimensionality [4] and that people can be described as self-centered or world-centered [2][3], but it does not show that one causes or predicts the other.
Self-centered cues show up in ordinary settings, but not as a dimensionality result
One study of memory and the built environment notes that associative memory relies on self-centered cues such as the orientation of your body [1]. That is a concrete example of a self-centered reference frame in action, and it shows the concept is grounded in everyday experience. It does not, however, measure the dimensionality of any learned space, so it adds a real-world illustration rather than evidence for the predictive link.
Sources used in this answer
Weaving Memories into Place: Exploring the relation between the built environment and associative memory
The paper examines associative memory in relation to the built environment and notes that it relies on self-centered cues such as body orientation, but it does not investigate the dimensionality of a learned space.
Experience and Reality
The paper discusses dimensions of experience and the contrast between self-centered and world-centered viewpoints, without clearly addressing the dimensionality of a learned space.
Know Yourself, Grow Yourself
The paper portrays lifelong learning as growth through awareness of dimensions such as perception and emotion, and contrasts self-centered with world-centered perspectives.
Data Modeling of Cognitive Structure in Physiotherapy Students Learning Gross Anatomy
The paper uses multidimensional scaling and Pathfinder networks to model the dimensionality of physiotherapy students' cognitive structures in learning gross anatomy, but does not address self-centered or world-centered reference frames.