Multidimensional Information Alignment
from Human Gestation to Age 3

Structural Isomorphism among Neurodevelopment, the Gut-Brain Axis,
the Maternal-Infant Connection, and Non-Scientific Narratives

I Introduction: How One Question Led to Four Timelines
How One Question Led to Four Timelines

The starting point of this paper is a simple question: “Why do humans have no memories from before age three?” This phenomenon — known as infantile amnesia — has long been attributed to neuroscientific explanations such as the immaturity of the hippocampus and the absence of a fully developed language system.

Yet when we followed this question layer by layer — from synaptic pruning to the gut-brain axis, from maternal microbial metabolites to the weaning reaction, from prenatal memory to theories of consciousness in non-scientific narratives — four independent factual timelines gradually surfaced and, without any anticipatory framing on our part, spontaneously revealed a striking structural alignment.

This paper records that process as it occurred. We emphasize: this is not a causal argument, not an endorsement of one side by another, but a “fact alignment” — arranging multiple independent data sources along a timeline and observing the degree of structural consistency between them.

II Dimension I: The Neural Explosion and Pruning
Dimension I: The Neural Explosion and Pruning

2.1 Gestation: From Zero to a Hundred Billion

At day 18–19 post-fertilization, the neural plate forms, marking the earliest appearance of neural tissue. By day 26, the caudal end of the neural tube begins to enlarge and the primordial human brain is born. Within the first 19 weeks of pregnancy, most brain cells have already been generated, at a peak rate of 500,000 neurons per minute. All brain neurons originate from the ectoderm and differentiate by way of the neural tube.

2.2 Birth to Age 3: Overconnection and Unconsolidation

At birth, the number of neurons is largely fixed (approximately 86 billion), but the synaptic network between them is under frantic construction. Before age three, each neuron produces on average around 15,000 synapses — far exceeding adult levels. This stage follows a “use it or lose it” principle: frequently used pathways are preserved; underused ones are pruned.

The brain as a whole is in “construction mode” rather than “running mode.” The synaptic network is extremely unstable and is continuously overwritten. A 2025 study published in Science (Yale/Columbia) further confirms that the posterior hippocampus of one-year-old infants can already encode episodic memories, but the prefrontal cortex responsible for retrieval is not yet online. Early memories are not deleted — they are “locked away.”

III Dimension II: The Gut-Brain Axis System Switch
Dimension II: The Gut-Brain Axis System Switch

3.1 Gestation: Remote Control by the Maternal Body

A landmark 2020 Nature study confirmed that the maternal gut microbiota modulates fetal brain development prenatally via microbial metabolites. Depleting the gut microbiota of pregnant mice produced defects in fetal thalamocortical axon development; selectively restoring the maternal microbiota prevented these abnormalities. Maternal microbial metabolites (short-chain fatty acids and others) cross the placental barrier into fetal circulation, participating in the formation of the blood-brain barrier and in the development of innate immunity.

The fetus’s own gut begins to “rehearse” operation during late gestation through swallowing amniotic fluid. Studies have found an overlap of microbial species between meconium and amniotic fluid, suggesting that the fetus may acquire its first maternally derived microbes in utero by swallowing amniotic fluid. The concept of a “maternal gut — placenta — fetal brain axis” has been formally proposed.

3.2 Birth to Age 3: From Proxy to Independence

After birth, the infant’s gut microbiota is in a highly unstable initialization phase. Microbes, antibodies, short-chain fatty acids, hormones, and other signaling molecules carried in breastmilk sustain a biochemical communication channel between mother and infant. Maturation of the gut microbiota occurs over the first 2–3 years of life, coinciding precisely with the critical window of early neurodevelopment. Over the first 12 months after birth, the variability of the microbial community both within and between individuals reaches its maximum, stabilizing only around age 3.

IV Dimension III: Establishment and Severance of the Maternal-Infant Connection
Dimension III: Establishment and Severance of the Maternal-Infant Connection

Integrating Dimensions I and II, we can identify a complete maternal-infant information chain:

A 2025 Cell Stem Cell study further confirms that the maternal gut microbiota releases metabolites such as butyrate and arginine, which activate the mTOR signaling pathway in fetal stem cells, thereby remotely regulating the behavior of fetal neural and intestinal stem cells. The microbiota does not “feed” the fetus so much as “direct” fetal cell fate through specific metabolic signaling molecules.

This means that weaning is not merely a switch of nutritional source — it is also the interruption of a signaling system. Information encoded in the old system (the maternal encoding format) may become unreadable by the new system (the independent encoding format).

V Dimension IV: Structural Descriptions in Non-Scientific Narratives
Dimension IV: Structural Descriptions in Non-Scientific Narratives

The factual data collected in this dimension come from three independent sources, which have no communication, coordination, or motivational linkage with the preceding three dimensions. This paper does not judge their truth or falsity; it includes them solely as independent factual data sources participating in the timeline alignment.

5.1 Prenatal Memory Research

A survey by the Japanese obstetrician Akira Ikegawa of 1,620 children with a mean age of 4 years found that 33% of them reported having prenatal memories and 21% could describe their own birth process. The expressible window for such memories is concentrated at ages 2–3 and vanishes rapidly after age 6. An independent survey by Professor Fumihiko Iida of Fukushima University in Japan reported similar proportions (53% prenatal memories, 41% birth memories). The Division of Perceptual Studies at the University of Virginia is also systematically collecting such cases.

5.2 The “Reborn Persons” Archive

The Tongdao Dong Autonomous County in Hunan, China, hosts extensive community-level records of “reborn persons” (individuals who claim to retain memories of a previous life). Over several decades, Ian Stevenson and Jim Tucker of the University of Virginia have gathered more than 2,500 worldwide cases of reincarnation-type memories. Such memories exhibit the same expressible window — concentrated at ages 2–6 — and are largely forgotten thereafter.

5.3 The Tibetan Buddhist Bardo Theory

The bardo theory describes a process in which consciousness leaves one body, passes through an intermediate state, enters the next embryo (rebirth / implantation), gradually becomes “stabilized,” and older memories fade away as the new body grows. This narrative system was codified about a thousand years before the birth of modern neuroscience and has no informational exchange with contemporary scientific researchers.

VI Four-Dimensional Timeline Alignment
Four-Dimensional Timeline Alignment

Arranging the facts from the four dimensions along a common timeline produces the following isomorphism:

VII An Untested Hypothesis
An Untested Hypothesis

Building on the alignment above, this paper proposes a testable hypothetical framework (not a conclusion):

The memory system in humans aged 0–3 may comprise two stages. The first stage (“maternal-connection memory”) runs from late gestation to pre-weaning: the infant encodes memories within the biochemical environment of the maternal-infant symbiotic system, and the encoding format depends on neural-network states modulated in part by maternal microbial metabolites. The second stage (“independent autonomous memory”) begins after weaning: as the gut-brain axis becomes independent, the microbiota stabilizes, synaptic pruning is completed, and the prefrontal cortex comes online, a completely new memory-encoding system is established.

Memories in the old encoding format are not “deleted” (as Science 2025 has confirmed) but cannot be retrieved because their encoding format is incompatible with the new system — much like files from an older operating system that cannot be opened on a new one.

Testable prediction: If this hypothesis holds, a correlation should exist between the timing of weaning and “the earliest age of retrievable memory.” Children who are weaned later may retain earlier memories; those weaned earlier may lose infant-period memories sooner. Combined with gut-microbiota transition data, this is a direction amenable to experimental design.

VIII Methodological Note: “Fact Alignment” Is Not “Fact Judgment”
Methodological Note: “Fact Alignment” Is Not “Fact Judgment”

The method employed in this paper is “fact alignment.” It differs fundamentally from both traditional scientific argumentation and the preaching of a belief system:

Scientific argumentation pursues the establishment of causal relations and reproducible verification. Belief systems pursue the internal consistency of doctrine and the spiritual experience of believers. Fact alignment establishes no causality and propagates no doctrine — it merely arranges facts from multiple independent sources along some dimension (here, time) and then observes whether structural isomorphism emerges between them.

The result of alignment cannot prove any party to be “correct.” A Nature paper cannot prove the bardo theory, and the bardo theory cannot predict the timeline of synaptic pruning. But when two descriptive systems — entirely unaware of each other’s existence — trace almost the same curve while describing the same phenomenon, that “uninvited consistency” is itself a fact worth recording.

IX Conclusion
Conclusion

This paper records the full process that began with a single question — “Why do humans have no memories from before age three?” — proceeded through layered data retrieval, and ultimately found a structural timeline alignment across four independent dimensions.

We have not drawn a conclusion. What we have drawn is an observation: four independent description systems, with zero communication between them — developmental neuroscience, gut-brain axis research, the biology of the maternal-infant connection, and non-scientific narratives about consciousness and memory — present a high degree of structural isomorphism in describing the single matter of “how a person comes to be this person, from nothing.”

There is no known reason why such isomorphism “should” exist. And yet it does.

Recording that fact is the sole purpose of this paper.