Statistical Learning - the ability to unconsciously extract patterns from sequences of stimuli over time.



Role of statistical and rule learning + SRT 


  1. Sequence motor skill learning - teaching a fixed sequence - RT is measured. RT goes up if the sequence is learned, and then disrupted. 

  2. Many mammals can learn this task 

  3. This is procedural learning → this involves the cerebellum and the motor cortex and kind of all over the place  

  4. Schendan et al 2003., this is an fMRI study where you will be given a button box → they will either be told to repeat the sequence or will be given a random order. With the random orders, you see additional blood flow in areas like the medial temporal lobe 

    1. It found activation in the prefrontal cortex (PFC), basal ganglia, and medial temporal lobe, all of which are involved in implicit learning and memory.

    2. This study showed that participants implicitly learned the sequence’s structure without conscious awareness—just like how people pick up language patterns without formal instruction.

  5. Amnesia patients have damage in the medial temporal lobe and showcase difficulty in implicit learning in the SRT tasks which suggests that the MTL contributes t implicit sequence learning 


Basal Ganglia 

  1. Basal ganglia lesions, but not hippocampal lesions, impaired SRT learning in rats.


SRT task + artificial grammar - this gets to the hierarchy structures 


  1. Train phase: you are exposed to the grammar structure - 

  2. Test phase: is this grammatical, does it follow the rule? - all of the sentences are new but they follow the same grammar as the train phrase - so this is testing the ability to generalise a rule and apply it to novel situations

    1. You can use letters - classic situation, very easy to follow  

    2. You can use nonsense sounds 

    3. Visual stimuli 

      1. New but wrong 

      2. New but correct 

        1. Brocas area is activated for both 

  3. AGL in infants - seen in 7 month olds - with very simple grammar 

  4. AGL in non-humans - seen in finches and cotton-top tamarins  - can learn ABB vs ABA 

  5. Hierarchical processing in AGL - only humans can learn phrase structure grammar 

  6. fMRI of Hierarchical AGL - when you contrast the two grammars ( Finite state transition grammar ABn and “phrase structure” hierarchical grammar A nB n) 

    1. If you violated both grammars there is elicited left frontal gyrus activation 

    2. Violations of FSG (the simple one)  activated the left frontal operculum, a region associated with simple rule processing and sequential structures.

    3. Violations in PSG activated both the left frontal operculum and Broca’s area, suggesting that Broca’s area is particularly involved in processing complex, hierarchical structures.



Word segmentation 

  1. fMRI - how different brain regions respond to speech stimuli with varying properties (e.g., stressed vs. non-stressed syllables) and how this relates to decision-making performance in a language-learning task.

  2. Both Auditory Cortices Respond to Speech Properties

    1. The left and right auditory cortices (including the superior temporal gyrus, STG) were activated when participants listened to speech stimuli.

    2. These regions were sensitive to differences in speech stress patterns, which play a role in language learning.

  3. Left Superior Temporal Gyrus (STG) & Learning Performance

    1. Higher activation in the left STG (a key area for processing speech sounds) was linked to better performance in a decision-making task.

    2. This suggests that individuals who had stronger neural responses in this region were better at recognizing statistical patterns in speech (e.g., identifying word boundaries in continuous speech).

  4. Visual word segmentation fMRI Turk-Brown et al 2008., 


Right StriatumImplicit Learning & Pattern Detection

  • The striatum is involved in habit formation and implicit learning.

  • The right striatum helps recognize statistical patterns (e.g., detecting repeating visual sequences).

  • Plays a role in reinforcement learning—helping the brain unconsciously learn what is predictable.

Left HippocampusMemory & Associative Learning

  • The hippocampus is critical for forming new memories and linking sequences together.

  • The left hippocampus likely helps in binding visual word fragments into meaningful units.

  • Supports long-term retention of segmented words.

Lateral Occipital LobesVisual Object Processing

  • The lateral occipital lobes process complex visual features like shapes and letters.

  • Helps in recognizing repeating visual patterns that form words.

Left Ventral Occipito-Temporal Cortex (vOTC)Visual Word Recognition

  • The vOTC is crucial for processing written words and letter sequences.

  • Includes the visual word form area (VWFA), which is specialized for reading.

  • This region may help convert visual patterns into readable units during segmentation.


Broca’s area 

  1. Petersson (2004) show Broca’s area also plays a role in general rule learning, sequence processing, and structured cognition—not just natural language.

  2. Violations in PSG activated both the left frontal operculum and Broca’s area, suggesting that Broca’s area is particularly involved in processing complex, hierarchical structures.

  3. Brocas type of aphasia do not do well in AGL 

  4. When you use non-linguistic stimuli and they had matched non-verbal iq and they still did worse 

  5. People with lesion to Broca’s area have demonstrated complex behavior. • Increasing evidence for other (more posterior) areas involved in complex cognition.

In this context, "complex behavior" refers to the unexpected or varied cognitive abilities that patients with Broca’s area lesions can still exhibit, despite traditional views that Broca’s area is essential for language and structured cognition.