Temporal coding carries more stable cortical visual representations than firing rate over time

Stably representing recurring visual scenes is crucial for behavior. However, previous studies report varying degrees of gradual neural activity changes over time in slow dynamic (1-5 seconds) firing rate code. Here we show that temporal codes, which capture structures in visually evoked fast (tens of milliseconds) spiking patterns, support the stability of visual representations. We tracked the spiking responses of the same visual cortical populations in male mice for 15 consecutive days using custom-developed, large-scale, ultraflexible electrode arrays. Across various stimuli, neurons exhibited different day-to-day stability in their firing rate-based tuning. The across day stability correlated with tuning reliability. Notably, temporal codes increased single neuron tuning stability, especially for less reliable neurons. Temporal coding further improved population representation discriminability and decoding accuracy. The stability of temporal codes was more correlated with network functional connectivity than rate coding. Thus, temporal coding may be essential in ensuring consistent sensory experiences over time.