Charan Ranganath: Human Memory, Imagination, Déjà Vu, and False Memories
Memory can feel like an archive, but neuroscientist and psychologist Charan Ranganath describes something far more dynamic. In this conversation with Lex Fridman, he explains that remembering is an act of reconstruction: fragments of experience are selected, combined with existing knowledge, and interpreted from the perspective of the present.
That flexibility allows memory to support imagination, learning, identity, and planning. It also makes recollection vulnerable to distortion. Across their discussion, Ranganath connects the biology of memory with practical questions about attention, childhood, sleep, heartbreak, misinformation, artificial intelligence, and the stories people tell about themselves.
The Experiencing Self and the Remembering Self
What happens during an experience is not identical to what remains afterward. Ranganath distinguishes the person living through an event from the person who later reconstructs it. Memory favors certain moments—especially those marked by surprise, emotion, change, or personal significance—while leaving much of the continuous experience behind.
This selectivity affects future decisions. People often choose whether to repeat a vacation, relationship, or challenge based on a compressed retrospective account rather than the complete experience. That account may be imperfect, yet it provides continuity and helps the mind extract lessons from the past.
Ranganath argues that this gives people some influence over how experiences become part of their identity. An unpleasant event can eventually be understood as meaningful, funny, or transformative without denying that it was painful. Difficult experiences shared with others can also become powerful sources of connection.
The self, from this perspective, is not a fixed entity stored somewhere in the brain. It develops as memories are continually interpreted alongside social relationships, changing goals, and new experiences. Adolescence is particularly influential because rapidly developing cognitive abilities, peer relationships, and formative events converge during that period.
How the Brain Creates—and Loses—Memories
Early childhood demonstrates that experience and durable autobiographical memory are different things. Ranganath explains that the hippocampus, which is central to episodic memory, continues developing during the first years of life. At the same time, the neocortex is changing rapidly, and a child’s first-person sense of self is still emerging. Together, these conditions help explain infantile amnesia.
Children may retain fewer detailed episodes than adults, but that limitation may support exploration. Their curiosity and tolerance for surprise encourage them to investigate information without filtering everything through rigid goals. As the prefrontal cortex matures through adolescence and early adulthood, behavior becomes more organized around plans, priorities, and self-control.
At the other end of life, forming new episodic memories may become harder. Yet Ranganath places this change in a social context: older adults preserve accumulated knowledge, practices, and cultural traditions. He points to grandmother-led orca groups as an example of how memory can benefit an entire community rather than only the individual carrying it.
Forgetting is not simply the erasure of a discrete file. Memories are represented across interacting populations of neurons, and retrieval can fail for several reasons. Similar experiences may compete, a memory may never receive the biological reinforcement needed for long-term retention, or the right retrieval cue may be absent.
Ranganath differentiates among several interacting functions:
Working memory keeps information temporarily available and directs its use.
Episodic memory represents particular events situated in a context.
Semantic memory supplies generalized knowledge about the world.
The brain combines these resources to build an internal model of what is happening. When reality departs from that model—through uncertainty, surprise, or a meaningful transition—the brain is more likely to encode a new episode. Event boundaries therefore act as important points for both storing and retrieving information.
Attention, Testing, and Practical Memory Training
Memory begins with selection. Neuromodulators, including dopamine and serotonin, help the brain prioritize what appears valuable, while attention determines which parts of an experience receive deeper processing. Ranganath presents expertise partly as trained attention: an expert notices patterns and distinctions that a novice has not learned to recognize.
This suggests that improving memory is less about expanding a generic storage capacity than about encoding and retrieving information more effectively. Memory competitors demonstrate the power of deliberate structure. With the method of loci, for example, a person imagines placing items along a familiar route and later retrieves them by mentally revisiting each location.
Other effective strategies share the same principles. Distinctive images create stronger cues, stories organize separate facts into meaningful relationships, and songs impose rhythm and sequence. Arbitrary information such as names becomes easier to recall when it is connected to a visual feature, location, or existing association.
Ranganath also emphasizes spaced repetition and retrieval practice. Repeated exposure distributed over time strengthens long-term access, while testing forces the learner to discover where recall fails. The struggle to retrieve is useful because it trains the memory system more actively than passive rereading.
These ideas translate into a straightforward learning approach:
Give the material undivided attention.
Connect it to knowledge you already possess.
Organize it into a meaningful structure.
Attempt recall without looking at the answer.
Revisit it after progressively longer intervals.
Multitasking works against this process. Constant switching introduces artificial boundaries into thought, burdens executive control, and makes it harder to integrate information over time. Ranganath notes that people with ADHD may remember highly engaging subjects exceptionally well while struggling when attention is not naturally captured. External structure and designated periods for specific tasks can help protect focus.
Why Imagination and Memory Are So Closely Linked
Imagination does not operate independently of memory. Ranganath describes it as the recombination of remembered components into situations that have not occurred. The hippocampus and default mode network participate in both recollection and imagined scenarios, helping the brain assemble people, places, goals, and possible outcomes.
Evidence from people with amnesia reinforces this connection. When access to episodic memory is severely impaired, constructing rich hypothetical scenes can also become difficult. The brain appears to preserve reusable elements of experience rather than complete recordings, allowing old material to support future simulations.
The same mechanism explains why remembering can be inventive. Every recollection integrates fragments with inference, knowledge, and current expectations. Imagining an event repeatedly may therefore make it feel increasingly familiar, blurring the distinction between something that happened and something merely considered.
This is not an accidental flaw. A perfectly rigid memory system would be less capable of generalizing, planning, and creating. The risk of distortion is inseparable from the flexibility that makes human thought adaptive.
Déjà Vu, False Memories, and Social Contagion
Déjà vu offers an especially vivid example of familiarity separating from accurate recollection. Ranganath notes that stimulation of areas in the temporal lobe, including regions near the hippocampus and amygdala, can produce déjà vu-like experiences. One possible explanation is that a present scene partially matches a stored pattern strongly enough to trigger familiarity, even though the person cannot retrieve a genuine prior episode.
False memories extend this problem beyond a fleeting sensation. Because recollection combines observed details with assumptions and later information, suggestion can become integrated into an event. The resulting memory may feel sincere and detailed despite being inaccurate.
The consequences become especially serious during coercive interrogation. According to Ranganath, stress, authoritative pressure, repeated suggestions, and plausible details can lead people to confess to acts they did not commit—and sometimes to develop apparent memories around the supplied narrative. Objective evidence has exposed such cases, showing that confidence alone cannot establish accuracy.
Memory distortion also operates at the group level. Communities construct shared accounts that reinforce belonging, but selective repetition can favor information consistent with the group’s identity. Social media can accelerate that process by creating clusters in which similar claims circulate without meaningful correction.
Ranganath suggests that diverse groups can form a more complete account when disagreement occurs within a climate of mutual respect. Different perspectives reveal omissions and challenge convenient reconstructions. Collective memory becomes more reliable when people can revise their beliefs without treating revision as betrayal.
Sleep, Time, and the Emotional Past
During sleep, patterns of hippocampal activity can resemble sequences produced during waking experience. Ranganath discusses this replay as part of the process by which new memories stabilize and become connected with older knowledge. Sleep does more than preserve individual events; it may help the mind discover relationships among them.
Memory also shapes the subjective passage of time. Looking backward, years can appear compressed because only a limited number of distinctive events remain accessible. Yet during an uneventful interval, time may feel slow because attention is directed toward its passage. Change makes lived time busy while also giving retrospective time more landmarks.
Heartbreak illustrates how emotion, memory, and identity interact. The end of a relationship combines intense feeling with major changes in daily life, making it highly memorable. Therapy can help a person reinterpret those memories—not by pretending the loss was pleasant, but by integrating what it revealed about love, vulnerability, and growth.
Nostalgia can similarly comfort or constrain. Recalling valued moments may improve mood and restore a sense of connection. But an idealized past can also narrow perception when its complexity is edited away. Ranganath’s broader point is that healthy remembering requires both appreciation and a willingness to update the story.
What Human Memory Can Teach AI
Ranganath and Fridman also consider whether artificial intelligence could benefit from something resembling episodic memory. Semantic knowledge gives a system generalized patterns, while episodic memory would allow it to learn flexibly from individual events. Combining the two is powerful in humans, but difficult to reproduce because an episode is embedded in goals, emotions, sensory context, relationships, and biological motivation.
Attention presents another challenge. Humans do not process every available detail equally; they move their eyes, gather information over time, and concentrate on what their internal model considers important. This efficiency also produces characteristic mistakes, including missed changes when expectations become too strong.
Machines have different sensors, objectives, and failure modes. Ranganath argues that reproducing human intelligence would require more than matching correct answers. A faithful model would also need to exhibit the distinctive biases and errors created by human perception, attention, and memory.
Brain–computer interfaces make these questions ethically immediate. As technologies become better at decoding neural signals or supporting speech, they could offer profound clinical benefits. They may also threaten privacy if internal intentions become increasingly accessible. Ranganath emphasizes the importance of transparency and regulation before technical capability outruns protections for freedom of thought.
Memory Is a Tool for Meaning, Not a Record of Reality
The conversation’s central lesson is that memory should not be judged by the standards of a camera. Its purpose is not to preserve every detail. It identifies what matters, builds models, supports prediction, and lets people revisit experience from new perspectives.
That constructive power makes memory both unreliable and indispensable. It enables imagination, creativity, social knowledge, and personal growth while leaving people open to suggestion, misplaced familiarity, and collective distortion. Understanding this tension can make us better learners—and more careful interpreters of our own certainty.
