Scientific Context
Ultima Memoria (UM) is situated within a scientific frame of reference that touches upon several central questions concerning human memory and cognitive performance. At the heart of the Ultima Memoria (UM) system are four interrelated dimensions of memory:
- the capacity of the memory,
- the speed of the memory,
- the form of expression and precision of the memory,
- long-term memory.
The analysis focused both on their functional interaction and on their measurability and modifiability.
Ultima Memoria (UM) is the first to treat these four dimensions of memory not as separate individual abilities, but as functionally interrelated components of a unified performance structure. A large memory capacity reveals its full practical value only when information is also rapidly accessible, precisely reproducible, and reliably available over the long term. It is this very interrelationship that forms the conceptual foundation of the Ultima Memoria (UM) system.
By contrast, existing research has, as far as can be determined, thus far examined these relationships only in isolated partial aspects. An integrated consideration of the capacity of the memory, the speed of the memory, the form of expression and precision of the memory, and long-term memory in the manner highlighted by Ultima Memoria (UM) is not yet discernible.
The selected research presented below addresses key partial aspects of this broader relationship. It considers memory capacity and processing speed; precision, quality of recall, and long-term availability; retrieval processes and durable retention; the relationship between memory, attention, and general cognitive performance; as well as neuroplasticity and the capacity of the adult brain to change. In addition, studies on the role of Artificial Intelligence (AI), technological support, and independent cognitive performance are included.
A |The Four Ultima Memoria (UM) Dimensions of Memory as an Integrated Performance Structure
Ultima Memoria (UM) structures, measures, and precisely defines memory performance within a four-dimensional framework:
- the capacity of the memory
- the speed of the memory
- the form of expression and precision of the memory
- long-term memory
B |Research on Memory Capacity, Speed, and Their Relationship
The relevant studies describe memory capacity as markedly limited. While Nelson Cowan identifies a central short-term memory capacity averaging approximately four chunks, the study by Rouder et al. calculates, within its model of visual working memory, an overall capacity of 3.35 information units (IU).
These studies thus demonstrate that the capacity of the memory can be empirically captured and measured. At the same time, they delineate an order of magnitude that Ultima Memoria (UM), according to its own measurement and development logic, substantially exceeds. Whereas the studies cited describe existing capacity limits, Ultima Memoria (UM) is directed towards the measurable expansion of the capacity of the memory. Moreover, Ultima Memoria (UM) situates capacity and speed within a broader system that also encompasses the form of expression and precision of the memory, as well as long-term memory.
Key Finding
Cowan revisits the classic claim that short-term memory can hold “7 ± 2” units of information (IU) and argues that, under more carefully controlled conditions, the more plausible estimate is closer to three to five chunks. His study thus offers a more precise account of the amount of information that can be maintained in an immediately accessible state over short periods.
Key Finding
Rouder et al. test models that assume a fixed upper limit to visual working memory. For the task format examined, their findings clearly favour a discrete fixed-capacity account: visual working memory appears able to retain only a limited number of mutually distinguishable units of information at any one time. On the basis of the underlying model, the average total capacity is estimated at 3.35 information units (IU).
Key Finding
The study explicitly tests whether memory speed and memory capacity are linked by a general relationship. More specifically, it asks whether the amount of information that can be held in short-term availability depends on how quickly those contents can be searched within memory. The findings, however, provide no evidence for a general, domain-independent law by which memory speed reliably accounts for capacity. At the same time, the authors stress that their results do not conclusively rule out such a relationship.
C |Research on Precision, Long-Term Memory, and Quality of Recall
The studies that follow focus on two further core dimensions of Ultima Memoria (UM): the form of expression and precision of the memory, and long-term memory. Their concern is not simply whether information is retained, but the accuracy, richness of detail, and durability with which it remains available. Taken together, they show that memory performance cannot be reduced to a binary distinction between "remembered" and "not remembered". The precision of recall, the quality of retrieval, and the sustained availability of stored information are equally significant.
These studies therefore engage with important aspects of the Ultima Memoria (UM) framework. Ultima Memoria (UM), however, takes this perspective considerably further by placing these dimensions within a broader account of memory performance. Rather than considering precision and long-term memory in isolation, it relates them to the capacity of the memory and the speed of the memory. This integrated perspective highlights a central premise of the system: memory storage attains its full practical value only when information is not merely retained, but remains available in large volume, at speed, with precision in expression, and over the long term.
Key Finding
Brady et al. examined how many visual objects people can retain over the long term and how precisely they are able to recognise them at a later stage. Over the course of several hours, participants viewed 2,500 objects and were subsequently able not only to distinguish previously seen items from new ones, but also to identify subtle variations in detail with a high degree of reliability. The study thus shows that visual long-term memory can combine an exceptionally large storage capacity with remarkable fidelity of detail.
Key Finding
The study challenges the widely held assumption that long-term memory is inherently less precise than working memory. Using highly familiar objects, Miner et al. show that visual memories retained over long periods can, under suitable conditions, be recalled with exceptional accuracy. Their findings therefore bring the precision of long-term memory into sharper scientific focus.
Key Finding
Richter et al. separate three dimensions of episodic remembering that are often treated as one: successful retrieval, the precision of what is recalled, and the subjective vividness of the memory. Their findings show that memory performance cannot adequately be described in binary terms as either “remembered” or “not remembered”. The precision of retrieval, in particular, emerges as a distinct and measurable aspect of memory in its own right
D |Research on Retrieval, Long-Term Availability, and Cognitive Performance Development
The studies that follow focus primarily on long-term memory, while also addressing how securely learned material remains available over time and how reliably it can later be retrieved. The central issue is not the mere intake of information, but the conditions under which it is retained over extended periods and remains accessible when needed. A recurring finding is that durable retention can be strengthened both through retrieval and through the temporal spacing of review.
These studies illuminate an important part of the Ultima Memoria (UM) framework, though not the framework in its entirety. Ultima Memoria (UM) treats long-term availability as one of four interdependent dimensions of memory rather than as a stand-alone phenomenon. It also adopts a different point of emphasis.
Whereas the research considered here examines lasting retention in relation to retrieval practice and spaced repetition, Ultima Memoria (UM) focuses on the measurable development and enhancement of long-term memory within a standardised overall system. According to its own methodology, this system requires no repetition of the learning material or, where repetition is necessary, limits it to a checking read-through.
Key Finding
Roediger and Karpicke compared the effects of repeated testing with those of additional study on later memory performance. Further study initially led to stronger results on a final test administered after only a brief delay. When recall was assessed two days and one week later, however, the testing conditions showed a clear advantage. The study is widely regarded as a foundational contribution to research on the so-called testing effect, the finding that active retrieval can enhance long-term retention.
Key Finding
Cepeda et al. examined more than 1,350 participants to determine how the interval between study and review sessions influences later retention. Their findings show that the most effective spacing between learning episodes depends on the intended retention period. Intervals that proved optimal for shorter retention periods differed from those best suited to very long-term memory extending up to one year. The study thus demonstrates that the temporal organisation of learning can exert a substantial influence on long-term availability.
E |Research on Neuroplasticity and the Brain's Capacity for Change
Any account of the development of human memory and cognitive performance must begin from the premise that the brain is not a fixed or immutable system. Ultima Memoria (UM) has long proceeded from the view that human mental capabilities are fundamentally capable of adaptation and development. Research on neuroplasticity has provided important empirical support for this perspective. It shows that learning, training, and experience may be accompanied by measurable functional and structural changes in the adult brain.
Research on the formation of new neurons, and on the particular plasticity of newly generated nerve cells, further suggests that neural changeability reaches deeper than was long assumed. At the same time, the evidence must be carefully distinguished. Training-related structural changes in the adult human brain are empirically documented, whereas the findings cited here on the heightened plasticity of newly formed neurons derive from animal studies.
This body of research is significant because it lends empirical support to a central premise: cognitive performance is not fixed, but capable of development. Ultima Memoria (UM) carries this premise into a concrete model of memory development. It relates the broader principle of cognitive adaptability directly to memory performance and translates it into a systematic framework for measurement and development.
Key Finding
Draganski et al. asked whether learning a complex visuomotor skill could alter the structure of the adult human brain. After three months of juggling practice, participants showed measurable increases in grey matter in regions involved in processing complex visual motion. Some of these changes diminished once the training stopped. The study therefore provides clear evidence that learning and training can induce structural adaptations in the adult brain.
Key Finding
Ge et al. found that neurons generated in the adult brain pass through a brief maturational phase of unusually high synaptic plasticity. During this window, they show an enhanced capacity for long-term potentiation, a form of synaptic change closely associated with learning and memory. Although based on animal experiments, the study suggests that adult neurogenesis may contribute to functionally meaningful neural adaptability.
F |Research on Technology, Artificial Intelligence (AI), and Independent Cognitive Engagement
Technological systems, and Artificial Intelligence (AI) in particular, are reshaping the conditions under which people access, process, and use information. Research shows that generative Artificial Intelligence (AI) can accelerate workflows, lower barriers to information, and improve the quality of certain outcomes. At the same time, more recent studies suggest that using such systems to replace rather than support a person's own mental engagement may be associated with reduced cognitive activation and weaker recall of self-generated content.
This raises a question of fundamental importance. The value of technological capability does not lie in displacing human memory and thought, but in connecting productively with them. Ultima Memoria (UM) addresses precisely this human dimension. It strengthens the memory and cognitive capabilities required not merely to access information more quickly, but to retain it independently, retrieve it with precision, assess it critically, and apply it with reflection.
The research considered here brings this tension into focus. It points, on the one hand, to the supportive potential of Artificial Intelligence (AI), and, on the other, to the continuing importance of active human cognitive involvement. Ultima Memoria (UM) brings these perspectives together by understanding the human being not as a passive recipient of technologically prepared information, but as an active bearer of memory, judgement, and conscious application.
Key Finding
Noy and Zhang examined the use of ChatGPT in work-related writing tasks. In a study involving 444 university-educated professionals, the tool reduced average completion time by 40% while increasing the assessed quality of the results by 18%. Their findings indicate that generative Artificial Intelligence (AI) can produce substantial productivity gains in certain areas of knowledge work.
Key Finding
The study explored how essay writing supported by large language models affects both brain activity and observable performance. Participants completed the task in one of three conditions: with an LLM, with a search engine, or without external tools. The first three sessions involved 54 participants, while 18 returned for a fourth session. Among the three groups, those using an LLM showed the weakest measured neural connectivity and the poorest recall of their own written content. The authors interpret this as a possible indication that Artificial Intelligence (AI) use may carry cognitive costs when it substitutes for, rather than supports, independent mental engagement.
