Lost in time: rethinking duration estimation outside the brain.

Our perception of the passage of time can suffer from significant distortions as time flies when we are busy and drags when we are bored. A prominent mechanistic model proposes that this perceptual volatility reflects changes in the activity dynamics of distributed neuronal ensembles referred to as population clocks because they encode time. In this framework, time is understood similarly to space (both can be segmented in seconds or centimeters) and duration estimation is primarily internal (the brain tells time). Here, I challenge this framework from the angle of Bergson’s proposal that the inner experience of time is unlike space because it is ever-changing and indivisible (2 successive seconds are not experienced equivalently). Quantifying and communicating this inner experience requires its externalization and immobilization through distance measurements derived from stereotyped movements and spatial metaphors (“short/long” durations; time “flies/drags”), which explains the habit of thinking time like space. In support of Bergson’s proposal, humans and animals heavily rely on movements in a variety of duration estimation tasks and the neural underpinnings of duration estimates overlap those of motor control and spatial navigation. Thus, philosophical and empirical arguments question whether duration estimation is fundamentally internal. Rather than being explained by ad hoc changes in the speed of population clocks, the puzzle of the volatility of time perception might resolve itself by considering that living beings lack the ability to internally measure time, which they compensate through interactions with regularities afforded by the world and symbolic representation drawn from space.

Encoding time in neural dynamic regimes with distinct computational tradeoffs

Converging evidence suggests the brain encodes time in time-varying patterns of neural activity, including neural sequences, ramping activity, and complex dynamics. Temporal tasks that require producing the same time-dependent output patterns may have distinct computational requirements in regard to the need to exhibit temporal scaling or generalize to novel contexts. It is not known how neural circuits can both encode time and satisfy distinct computational and generalization requirements, it is also not known whether similar patterns of neural activity at the population level can emerge from distinctly different network configurations. To begin to answer these questions, we trained RNNs on two timing tasks based on behavioral studies. The tasks had different input structures but required producing identically timed output patterns. Using a novel framework we quantified whether RNNs encoded two intervals using either of three different timing strategies: scaling, absolute, or stimulus-specific dynamics. We found that similar neural dynamics for single intervals were associated with fundamentally different encoding strategies and network configurations. Critically, some regimes were better suited for generalization, categorical timing, or robustness to noise. Further analysis revealed different connection patterns underlying the different encoding strategies. Our results predict that apparently similar neural dynamic regimes at the population level can be produced through fundamentally different mechanisms—e.g., in regard to network connectivity and the role of excitatory and inhibitory neurons. We also predict that the task structure used in different experimental studies accounts for some of the experimentally observed variability in how networks encode time.

Time cell encoding in deep reinforcement learning agents depends on mnemonic demands

The representation of "what happened when" is central to encoding episodic and working memories. Recently discovered hippocampal time cells are theorized to provide the neural substrate for such representations by forming distinct sequences that both encode time elapsed and sensory content. However, little work has directly addressed to what extent cognitive demands and temporal structure of experimental tasks affect the emergence and informativeness of these temporal representations. Here, we trained deep reinforcement learning (DRL) agents on a simulated trial-unique nonmatch-to-location (TUNL) task, and analyzed the activities of artificial recurrent units using neuroscience-based methods. We show that, after training, representations resembling both time cells and ramping cells (whose activity increases or decreases monotonically over time) simultaneously emerged in the same population of recurrent units. Furthermore, with simulated variations of the TUNL task that controlled for (1) memory demands during the delay period and (2) the temporal structure of the episodes, we show that memory demands are necessary for the time cells to encode information about the sensory stimuli, while the temporal structure of the task only affected the encoding of "what" and "when" by time cells minimally. Our findings help to reconcile current discrepancies regarding the involvement of time cells in memory-encoding by providing a normative framework. Our modelling results also provide concrete experimental predictions for future studies.

Specific loss of neural sensitivity to interaural time difference of unmodulated noise stimuli following noise-induced hearing loss

Sensorineural hearing loss compromises perceptual abilities that arise from hearing with two ears, yet its effects on binaural aspects of neural responses are largely unknown. We found that, following severe hearing loss because of acoustic trauma, auditory midbrain neurons specifically lost the ability to encode time differences between the arrival of a broadband noise stimulus to the two ears, whereas the encoding of sound level differences between the two ears remained uncompromised.

Lithuanian temporal adverbials: position in the system of temporal expressions and a review of semantic research

Expression of time is an important field in linguistic research. In Lithuanian linguistics, the study of time has focused first and foremost on tense and aspect systems. Therefore, Lithuanian language has developed a big diversity of means to encode time, e.g. noun cases (vasarą ‘in summer’) and noun phrases (liepos vakarą ‘in July evening’), prepositional phrases (po darbo ‘after work’), adverbs (netrukus ‘soon’), subordinate clauses (kai nustos lyti ‘when it stops raining’). All of them despite their different grammatical status are suggested to be called temporal adverbials. ‘Adverbial’ is a quite new notion in Lithuanian linguistics, it was first presented in the studies of stance adverbials but is also convenient in the field of linguistic expression of time. The purpose of the paper is to demonstrate the necessity of distinguishing a semantic-functional class of temporal adverbials in Lithuanian linguistics and to show their position in the system of means that encode time. The second part of the paper presents a few semantic models of the temporal adverbials made by foreign linguists and the semantic research of Lithuanian temporal means that should be regarded as temporal adverbials.

The Effect of Language on Economic Behavior: Evidence from Savings Rates, Health Behaviors, and Retirement Assets

Languages differ widely in the ways they encode time. I test the hypothesis that the languages that grammatically associate the future and the present, foster future-oriented behavior. This prediction arises naturally when well-documented effects of language structure are merged with models of intertemporal choice. Empirically, I find that speakers of such languages: save more, retire with more wealth, smoke less, practice safer sex, and are less obese. This holds both across countries and within countries when comparing demographically similar native households. The evidence does not support the most obvious forms of common causation. I discuss implications for theories of intertemporal choice. (JEL D14, D83, E21, I12, J26, Z13)